Datasets:

ai-aerospace

/

ams_data_full_2000-2020

like 0

497 Approved for public release. OTR 2020- 00294 Invitations were sent out as a call for participation on the industry committee to meet on a regular basis and adjudicate all identified CRM items . These invitations were sent to a wide variety of spacecraft, payload, and launch vehicle contractors, as well as government and general interest participants, as requested by AIAA, who also requested representation from academia, but no such individuals could be identified as specifically concerned with the subject matter of this standard. Once invitations had been accepted by a quorum of the aforementioned categories, regular periodic teleconferences were set up to discuss the CRM items and help adjudicate each one. Based on a preponderance of participation on those teleconferences, the recognized MMA Committee on Standards included the following individuals and organizations: • Brian Gore The Aerospace Corporation (Co-Chair /Author ) • Leon Gurevich The Aerospac e Corporation (Co-Chair /Author ) • Mark Balzer Jet Propulsion Laborator y • Ed Boesiger Lockheed Martin Corporation • Ray McVey Raytheon (retired) • Mike Pollard Lockheed Martin Corporation • Brandan Robertson NASA Johnson Space Center • Adam Sexton Ball Corporation • Tim Woodard The Aerospace Corporation Invitations were tentatively accepted by individuals at Boeing, ULA, and Northrop Grumman, but their participation in the per iodic tel econferences was limited . From early 2018 through April 2019, on approximately 2- week intervals, the committee met and discussed each of the CRM items, as well as contributed their organizations’ concerns and additional CRM items to the list as the effort progressed. By the conclusion of the adjudication effort there were 136 Parsing items, 93 Testing items, and 117 Non- Testing (or “other”) items (3 46 total) that were dispositioned as either accepted or rejected in the new standard. The Most Significant Changes The three most significant changes in the updated MMA standard are the removal of the “shall, where practi cal” weighting factor, a newly re vamped testing section, and a significantly modified method to calculate force/torque margi n. These changes are described in greater detail in the following sections . Elimination of “ Shall, Where Practical” During the pre- committee phase of the document review, the standard was examined for all instances of the “shall, where practical” wei ghting level requirement . These instances were added as line items on a CRM tab labeled Parsing. Because of numerous progr am tailoring exercises where feedback was received from organizations unsure about how to treat these – are they real/actual requirements to be traced in a database or not? – the authors attempted to address each one on its own merit and either upgrade it to a strict “shall” requirement, or reduce it to a “should” type of guideline for recommended engineering practice. Each of these items was then discussed during the industry committee telecons, and a consensus was reached for each one. Overall, there were approximately 85 items designated at the “shall, where practical” weighting level in the 2005 version of the standard. Of those, only about 13 were upgraded to the “shall” weighting level, about 26 were downgraded to the “should, preferred, may” weighting level, 32 were absorbed and modified into the new testing section requirements, and 14 were removed entirely for being out of scope, not applicable, or covered by other requirements. One should note that a simple comparison count of the number of “shall” requir ements in the old vs. new standards is not meaningful, largely because of the reformatting of the testing section, in which many requirements are repeated at the unit, subsystem, and/or vehicle level of assembly, which was not delineated in that fash ion in the old version.

Document

AMS_2020.pdf

498 Approved for public release. OTR 2020- 00294 Testing Section During the previous update to the standard in 2005, AIAA deadlines limited the MMA Committee on Standards focus on the testing section. As a result, much of the previously existing material was still held over from the boiler plate language found in the (even then) decades -old processes and philosophies . With the current standard update, the testing area was the first area of emphasis to modernize, making it more usable and valuable to the engineers who employ it. A realization was made by the authors that many of the same programs that have this MMA standard levied as a set of requirements also have other testing standards levied at the same time, such as those listed in References 1- 4. Since the authors (and a compendium of MMA standard users on government programs ) were most familiar with References 1 and 2, many of the same outline, structure, and philosophical elements found in those testing standards were duplicated in the new MMA standard where it made logical sense. For example, the previous version of the MMA standard did not give any guidance or direction regarding which tests were intended to be performed at each level of as sembly . To remedy this, the new MMA testing section mimic s that of Reference 2, with separate Unit Level, Subsystem Level, and Vehicle Level test subsections . It was noted that one of the more referenced features in the testing standards is a table or matrix that indicates what testing should be done at which level . In order to assist the user even further in this area, a similar table was constructed for the MMA standard, and is shown in Figure 2. With this table, one can see at a glance the required test s (and those requiring a formal evaluation to be conducted to determine if they are required) for each level of assembly, as well as for the test type (qualification, proto- qualification, or acceptance) . Users of the standard will also notice that the slate of typical MMA tests have been broken down into broad categories such as P erforma nce tests, E nvironmental Exposure tests, and “Special” tests. Another omission in previous versions of MMA standards is any guidance regarding the flow of a testing program, and in what order these required tests should be performed. Many programmatic, personnel, or resource limitations can dictate the order of some testing flows, but there was no default or recommended order. As such, recommended test flows were added to the standard, with an example shown in Figure 3 that illustrat es a typical test flow sequence one might employ for an MMA at the Unit Level . Additional, similar test sequences are provided in the standard for Subsystem and Vehicle L evels of assembly as well. These sequences provide not only a default recommended order by which to conduct a test program, but also provide respective reference paragraphs in the document for more detail on each particular test. Th e reference paragraphs atte mpt to provide test -specific requireme nts as well as bac kground information, such as the purpose o f the test or other nuances that may help perform or set up a particular test . A simple color - coding scheme also identifies which tests may only be applicable for qualification or acceptance programs specifically.

Document

AMS_2020.pdf

499 Approved for public release. OTR 2020- 00294 Figure 2. Testing requirements matrix in new MMA standard

Document

AMS_2020.pdf

500 Approved for public release. OTR 2020- 00294 Figure 3: Typical MMA test sequence at unit level Torque/Force Margin Approach In this discussion , references to “torque margin” also apply to linear motion, with “force” simply replacing “torque.” This subject may be the biggest change in the new MMA standard, and one that will likely take the most adjustment with respect to the traditional approach for experienced MMA engineers . This new approach to calculating torque margin was first widely discussed around the time of the previous AIAA MMA standard release, but the aforementioned AIAA deadline prevented any further committee discussions on the matter. Furthermore, this approach has already been published in its basic form in NASA’s MMA standard, Reference 5, but there were some modifications and accommo dations made in this AIAA MMA Standard update. The basic premise was that there are at least three different types of resistive forces that MMAs encounter – those that are relatively predic table and repeatable , those that are variable and harder to predict initially or over life, and those that are purely the result of an induced acceleration. Examples of the first kind, or “fixed” resistances, are a return or fail -safe spring, fluid pressure in a valve, etc . Examples of the second kind, or “variable” resistances, are coulomb fri ction, cable harness bending stiffness, etc . Examples of the third kind, or “acceleration- depende nt” resistances, are those produced by rotary motion, launch vehicle accel eration, etc. In the new torque margin calculation approach, which resembles the method in which stress or mass margins are calculated, each resistive torque carries with it a different uncertainty factor. Traditionally, with a torque margin requirement of 100%, all resistive torques were assigned an uncertainty factor of 2.0. In the new approach, the uncertainty factor is a function of the “variable” or “fixed” nature as described above, as well as of the method by which it s value is obtained, such as by analysis, testing on an engineering model or qualification article , or testing on an actual flight unit. Figure 4 shows the minimum torque/force margin factors used in this new approach.

Document

AMS_2020.pdf

501 Approved for public release. OTR 2020- 00294 Figure 4: Minimum Torque/Force Margin Factors Before discussing the new calculation method and equation(s), it is also worthwhile to mention that there are now five different types of torque/force margins identified , with the requirements of each needing to be satisfied as appropriate for a given MMA or subsystem: • Static Torque/Force Margin – a measure of the excess torque available to overcome resi stance to impending motion • Constant Velocity Torque/Force Margin – a measure of the excess torque available to maintain motion • Holding Torque/Force Margin – a measure of the excess torque available to maintain position in the presence of external disturbances • Dynamic Torque/Force Margin – a measure of the excess torque available to accelerate a body by a given amount • Stepper Motor Margin – a measure of the excess pull -in torque at the drive rate available to overcome friction loads seen at the motor The ne w single equation used to calculate the torque/force margin is shown in Equation 1: Torque Margin =Τavail ΣΚfixΤfix +ΣΚvarΤvar+ΣΚ accΤacc −1 (1) where - Tavail is the minimum available torque generated by the driving or holding component (e.g., spring, motor). - Tfix terms are the individual maximum “fixed” resisting torques that are not strongly influenced by effects of friction, temperature, cycles, etc. (e.g., motor detent torque, vehicle maneuver -induced torque, return spring torque, unbalanced pressure load limited by relief mechanisms). - Tvar terms are the individual maximum “variable” resisting torques whose values may change with environmental conditions a nd cycles (e.g., friction torque, wire harness torque due to flexing or long term set). - Tacc is the torque required to achieve the specified acceleration of the driven component. - Kfix, Kvar, and K acc are the fixed, variable, and acceleration torque/force margin factors applied to each individual resisting torque/force term in Equation 1 prior to summation. As mentioned earlier, torque/force margins are now calculated in the same manner as stress, strength, and mass margins . As such, the requirement for st atic torque margin is no longer 100%, but that it be positive (> 0) , after the appropriate uncertainty factors are applied. All terms in Equation 1 may not be used in all MMA applications, and further guidance for each type of margin is provided in the standard.

Document

AMS_2020.pdf

502 Approved for public release. OTR 2020- 00294 For stepper motor torque margin, engineers are given a choice to use either a step stability analysis or a pull-in torque margin analysis . The step stability analysis is stated as preferred, but the pull -in torque method is also acceptable, only when certain conditions are met . Equation 2 shows the pull -in torque margin calculation method. Pull-in torque margin = �� Pull−in torque at drive rate Kvar ∗ Total friction load seen by motor �−1� (2) When choosing the step stability analysis method instead, Equation 2 can be re- written to take the form of Equation 3 Pull-in torque at drive rate =[1+Pull-in Torque Margin ]∗[Total friction load seen by motor ] (3) Step stability is achieved when the left side of Equation 3 is great er than the right side. In the step stability analysis, the value of the [1+ Pull -in Torque Margin] term is increased until the motor goes unstable. The value from the last stable case is then used to calculate the stepper motor margin . Summary and Path F orward A multi -year effort has culminated with a finished draft of a newly updated AIAA Moving Mechanical Assemblies standard. The most significant changes have been discussed herein. These include the elimination of the “shall, where practi cal” requireme nt weighting level, a completely new testing section that mirrors other industry testing standards, and a more realistic approach to calculating torque/force margins . There are a host of other changes which were not highlighted in this paper, but which the authors and industry committee members believe will make a more robust and meaningful set of MMA requirements for space and launch vehicle programs in the future. As of the submission of this paper (January 2020), the draft document is making its way thr ough at least two approval processes, one by The Aerospace Corporation to replace the Mission Assurance Baseline callout of the previous MMA standard, and the other by the AIAA. It is anticipated that these will be completed sometime in 2020. References 1. SMC -S-016, Air Force Space Command, Space and Missile Syst ems Center Standard, “Test Requirements for Launch, Upper Stage, and Space Vehicles, ” September 2014. 2. TR-RS-2014- 00016, The Aerospace Corporation, Technical Report, “Test Requirements for Launch, Upper -Stage, and Space Vehicles ,” June 2014. 3. MIL-STD-1540 C, Department of Defense, Military Standard, “ Test Requirements for Launch, Upper - Stage, and Space Vehicles ,” September 1994. 4. GSFC -STD-7000, NASA Technical Standard, “General Environmental Verificati on Standard (GEVS) ,” April 2013. 5. NASA- STD-5017A, NASA Technical Standard, “Design and Development Requirements for Mechanisms,” July 2015. Acknowledgements The authors wish to acknowledge Rebecca McKenna of The Aerospace Corporation’s Enterprise Mission Engineering Department, whose unwavering positive support allowed a long- term, low -level effort to complete this work despite continually shifting and competing priorities.

Document

AMS_2020.pdf

503 Micro -Vibration Attenuation Using Novel Flexible Pivot Design Luc Blecha*, Yoël Puyol*, Simon Hayoz *, Martin Humphries** and Fabrice Rottmeier * Abstract Flexible pivots present numerous advantages such as no wear, no particle generation and predictable torque. In addition, the patented design presented in this paper adds high stiffness tunability, large angle , nearly constant radial stiffness over the enti re range of motion , and infinite life capability. A design optimization software has been developed to generate within a few hours custom flexible pivot designs matching specific application requirements such as pivot angle, torsional, radial, axial , bendi ng stiffness , maximal stress, and buckling factors. The flexible pivot design is an enabling technology for many applications and in particular for ultra- stable pointing mechanisms . A novel concept of an ultra- stable pointing mechanism using a flexible piv ot is presented. It is shown with a simple Nastran model that the micro -vibration impact on pointing mirror stability can be theoretically decreased by 3 to 8 orders of magnitude in the 50- 200 Hz frequency range, and by more than 9 orders of magnitude above 200 Hz in comparison with a design using ball bearings . The achieved pointing accuracy makes the need of a fine pointing mechanism unnecessary. The total mass, volume and costs can thus be drastically reduced in comparison to the existing s olution today on the market that uses coarse and fine pointing mechanisms . A fully functional and motorized breadboard has been built and showed full hemispherical pointing range. Large -Angle Flexible Pivot Design The principle of the Large- Angle Flexib le Pivot is based on controlled deformation of structures in pure bending within their elastic limit. The flexible pivot consists of two interface rings connected by flexible elements (Figure 1) . The first ring is the stator and the second is the rotor. A first set of flexible elements connect the stator ring to a central cylinder. These spokes are composed of a thin elongated blade connected to another smaller blade called T -bars as they are forming a T -shape . The T -bars are connected to the stator. From the central cylinder, connecting members connect the first central cylinder to an outer ring by means of another set of T -bars. From the outer ring, the structure is symmetrical . The outer ring is thus connected to another central ring on the rotor side by connecting members and T -bars. Finally, the second central ring is connected to the rotor ring by a set of spokes and a f ourth set of T -bars. This is a so - called two -stage design that can achieve a rotation of ±70° for infin ite life conditions and higher angular range with reduced life. * Almatech SA, Lausanne, Switzerland; luc.blecha@almatech.ch ** SpaceMech Ltd., Bristol, UK Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 20 20

Document

AMS_2020.pdf

504 Figure 1. Design Overview of the Almaflex The number of spokes, connecting members and T -bars can be adjusted to the need of the application. For example, if a larger rotation angle is needed, the blade can be positioned at 120 degrees instead of 90 degrees. The number of spokes and connecting members would thus be 3 instead of 4. The number of stages can also be increased which will increase the range of rotation al angle. The number of thin blades, and spokes can also be increased to raise the torsional stiffness. This is used in oscillating applications to reach high natural rotational eigenfrequencies. All these applications are documented in a world patent1. A) B) Figure 2: A) View of a 3 stage outside- inside design with parallel spokes and connecting members to increase stiffness B) a 3 stage outside- outside design for very large rotational angles.

Document

AMS_2020.pdf

505 The advantage of this unique flexible element configuration is that the sti ffness in each degree of freedom can be tuned almost independently from the other direction. In particular, the T -bars drive the radial stiffness whereas the total height of the blades drives the axial one. The rotation of the flexure is driven by spokes for approximately one half of the full range while the other half is provided by the connecting members. This decoupling feature is particularly useful for stiffness optimization considering in -orbit operations, launch locking, gravity sag adjustments and c an easily be verified via non- linear Finite Element analysis. The fully symmetrical design guarantees no geometrical center shift and, combined with T -bars compliance, a smooth thermo- elastic behavior. Custom Design for E ach Application The Almaflex design is a family of designs that can be optimized to specific performances. The Large Angle Flexible Pivot (LAFP) is the result of an optimization in response to the European Space Agency (ESA) CTP specifications. A software tool called FlexOptim has been developed to efficiently find an optimized design within the Almaflex design family that matches best the application requirements. Rotational angle, applied loads , and launch configuration are given as input to FlexOptim as well as stiffness requirements for radial, axial, torsional and bending cases. The optimization process starts with a baseline set of variables and associated boundaries def ining the initial geometry. A set of constants defining material properties and weighting factors are also input. Depending on the configuration during launch: locked or not, a set of design forces is retained for the optimization process. According to the requirements, variables and constants given as input, constraints are generated and cover geometry consistency, targeted stiffness , angle, stress limit and buckling factors. T he optimization process starts from the initial set of variables . The optimizati on process is done in two steps. Optimization loops are performed using analytical formulas that calculate axial, radial, bending and torsional stiffness as well as the stresses and buckling factors . Consistency of these formulas has been correlated with f inite element analysis during process development. The optimization aims at minimizing torsional stiffness, volume, and the deviation between targeted and computed rotation angles while maximizing buckling factors and stress performances. In a second step, a complete finite element loop is performed to evaluate the requirements. Correction factors between FE analysis and analytical approach are then computed and considered in the next analytical iterations. The optimal solution is reached when convergence of these correction factors is met. The generic flow logic of the optimization algorithm is shown in Figure 3. Figure 3: Optim ization methodology

Document

AMS_2020.pdf

506 Manufacturing constraints are considered by the specification maximal blade length, minimal blade thickness , and allowable range for blade length to height ratio. In addition, specified allowable stress and elastic modulus shall repres ent achievable values by the end product. To this end, a fatigue test campaign has been set up. Establishment of Wohler curve for titanium grade 5 is currently underway for an ESA CTP project. The preliminary results have shown that the LAFP maximal number of cycles is highly dependent on the manufacturing process (see reference 5 for further details). By selecting the adequate process, 309 million cycles could be achieved without failure. Once the fatigue test s are finis hed, failure probability analysis will be carried out based on the Wohler curves to identify the maximal safe operation of the LAFP. The optimization tool used in FlexOptim to minimize or maximize the objective functions is MIDACO2,3,4 (Mixed Integer Dis tributed Ant Colony Optimization). MIDACO is a metaheuristic optimizer developed by ESA based on the behavior of ants looking for food around their hill, so its basis of functioning is the exploration of the available space for the variables and the storage of the current best solution. The constraints are handled by using an oracle penalty method. As shown on Figure 4, different levels of exploration can be observed and finally a smaller area around the optimal solution is investigated. Figure 4: Example of optimization results from MIDACO Once the optimum geometry fulfils the customer’s requirements , the new set of geometry variables are sent automatically to a 3D design software that updates the design. A Large Family of Design s Based on the optimization method described above, different designs have been established for a variety of applications. Some of them have been selected to show the different possible design and their main characteristics are shown in Table 1 .

Document

AMS_2020.pdf

507 Table 1: Main characteristics of some selected applications Design identification D1 D2 D3 D4 Characteristics Unit Application Pointing Pointing Slow Scan Rapid scan Payload configuration Supported on both ends Supported on both ends Supported on both ends Supported on both ends Launch configuration Launch lock Launch lock Launch lock Launch lock Payload mass kg 2 4 1.2 3.0 Flex material Titanium PEEK Titanium Stainless Steel Rotation angle Degrees ±8.5 ±45 ±70 ±45 Torsional stiffness Nm/rad 0.33 0.45 0.98 27.3 Axial stiffness N/m 160000 103000 580000 592000 Radial stiffness N/m 160000 114000 132000 422000 Radial to torsional stiffness ratio Rad/m2 474800 252400 134000 15500 External diameter mm 47 52 100 172 Overall height mm 10 23 45 40 Flex mass kg 0.080 0.027 0.269 0.627 The different designs are shown in the Figure 5. Design D1 Design D2 Design D3 Design D4 Figure 5: Different achievable design listed in Table 1, represented respecting their relative size

Document

AMS_2020.pdf

508 PEEK material is used in design D2. This material is very interesting as it is relatively stiff, has good mechanical strength chemical stability , and very good fatigue behavior at the current testing stage. The fatigue stress to Young’s modulus ratio, which indicate s the ability of a material to deform without fatigue rupture, is 2 to 3 times higher than titanium. A very compact design with large angle is thus achievable, while maintaining high radial to torsional stiffness ratio. In addition, it is very light. Design D3 is a high -end application for large angle application. The total range of rotational angle is 140 degrees ( ± 70 degrees), while maintaining lifetime above 100 million cycles with applicable safety margin. The stress and torsional stiffness have been minimized. Center shift is also kept below 10 microns on the entire range of motion. More details on this design can be found in R eference 5. Design D4 has been developed for a high- frequency oscillatory application that uses the el astic energy as storage. In this application, a cog- free motor is exciting the first torsional mode of the flex with synchronized impulse, which compensates the material damping loss. At each oscillation, the range of motion is increased to reach a maximal oscil lation of 90 degrees pea k-to-peak (±90 degrees). As the energy stored in the movement is proportional to the flexure’s stiffness and to the angular range of motion squared , high stiffness and large angle shall be targeted. In this application, the total el astic energy is 8.4 J per flexure. It is worth noticing from Table 1 that the LAFP design optimization process is able to achieve very different design goals. For example, in design D1, the application required a relatively small rotational angle. The benefits of the Almaflex design family are the absence of tribological effects thus no wear nor particle generation, and a stable, predictable behavior all -along lifetime. In addition, one unique feature is that the radial stiffness is kept nearly constant during the full rotation. Only 17% stiffness drop is observed between the radial stiffness at 0° rotation and 70°. This feature is essential for micro -vibration isolation. Micro -Vibration – A Pointing Disturbance One of the applications of large angle flexible pivot is precision pointing mechanism, such as laser communication terminal, scanner, and flip mirrors. Pointing from a platform to a 1 m diameter tar get that is distant of 1000 km (Low Earth Orbit) requires a pointing accuracy and stability that is at least a fraction of 1 µrad, typically 0.1 µrad. Such stability requirement may apply to Earth observation satellite, and laser communicat ion between space and ground. For intersatellite laser communication link, the stability requirement is more stringent as the distance between satellites is generally larger, and diameters of target optics are smaller. To achieve a pointing stability of 0.1 µrad or smaller, understanding the micro -vibration effect on pointing stability is essential . There are sources of vibration noise on each platform . These sources are coming from different moveable parts such as a shutter, solar array drive, reaction wheels , and cryo -coolers. Each of these mechanisms has a micro -vibration signature. Reaction wheel s are one of the noisiest equipment. As shown in Figure 6 , reaction wheels typically generate low vibration noise at low frequenc ies and high noise at frequencies above 200 Hz.

Document

AMS_2020.pdf

509 Figure 6: Example of micro -vibration noise specification of reaction wheels for MTG On the other hand, c ryo-coolers generate most of their micro -vibration noise at frequencies around 50 Hz and less at higher frequencies . Note that the level of micro -vibration from cryo -coolers is generally smaller than those of r eaction wheel s, even at 50 Hz. Figure 7: Example of micro -vibration noise spectrum for cryo-coolers on MTG The different sources of micro- vibration add up at the pointing mechanism interface. The micro -vibrations are then transmitted to the mirror by the pointing mechanism which disturb the pointing stability. Translation vibration of the mirror are generally not a concern as it does not affect pointing stability. On the other hand, rotational vibration of the mirror has a direct effect on the pointing stability . The transfer function between the pointing mirror and the pointing mechanism’s interface to the spacecraft is thus key. In case of a stiff pointing mechanism using ball bearings, the structure trans fer function has a first amplification peak at its first structural eigenfrequency , which is typically between 140 Hz for spacecraft decoupling requirement and below 1000 Hz. Bel ow the first eigenfrequency, micro- vibration s are either transmitted to the mirror integrally for frequencies far below the peak or amplified for the ones near

Document

AMS_2020.pdf

510 resonance. Thus, cryo- cooler s micro -vibrations that have the highest excitation at frequencies between 50 and 100 Hz are integrally transmitted to the pointing mirror. In addition, high level micro -vibration noise from reaction wheel s may well coincide with the first structural eigenmode of the pointing mechanism, leading to huge amplification. The f irst eigenfrequency of the pointing mechanism shall thus be tuned to match the reaction wheel s quiet area. To illustrate this phenomenon on a classical series az imuth- elevation pointing mechanism, a simplified Nastran model was built (see Figure 8). Figure 8: Description of the simplified Nastran model The interface of the mechanism to the spacecraft is model ed by a single node and a CBUSH element. The CBUSH has the stiffness properties of the az imuth ball bearing with low rotational stiffness around Z axis. The azimuth fork connects the az imuth ball bearing to the upper frame. The elevation axis is connected to the upper frame by two CBUSH s representing the elevation axis ball bearings. The mirror Co G is located at the intersection of the azimuth and elevation axis. The beam section has been tuned to obtain a first frequency above 140 Hz, and near the quiet reaction wheel zone that is below 200 Hz and the CBUSH spring constant has been chosen to represent ball bearings stiffness . The total mass of the modeled mechanism is about 7 kg. The eigenfrequencies and modal effective mass of the simplified model of a pointing mechanism using ball bearing is shown in Table 2. The participating mass and inertia that are dominant for each mode has been highlighted in Table 2. Table 2: Ball bearing pointing mechanism eigenfrequencies and participating mass ID Freq TX TY TZ RX RY RZ Hz kg kg kg kg m2 kg m2 kg m2 1 0.2 -2.74E -05 -1.33E -05 1.95E -16 1.10E -08 1.46E -10 -2.13E -01 2 4.3 1.62E -15 -3.55E -05 -8.11E -10 .521E-01 -2.45E -17 -.116E-12 3 163.4 -3.29E -10 2.59E+00 .278E-05 -4.09E -03 -.242E-10 -2.54E -12 4 179.9 1.15E+00 3.98E -10 -2.81E -05 1.23E -11 1.68E -01 -1.78E -12 5 499.6 -2.33E+00 6.44E -12 -2.46E -05 2.31E -11 .845E-01 .544E-12 6 779.9 -.974E-05 -1.54E -05 2.57E+00 -1.54E -05 1.51E -06 9.89E -18

Document

AMS_2020.pdf

511 The first two modes are the ones associated with the az imuth and elevation rotation ax es. As these modes are controlled by the elevation and azimuth motors, they are not a concern in the micro -vibration analysis. The 3rd mode is a translational mode in Y direction and takes place at 163 Hz. The 4th and 5th modes are due to the bending of the azimuth fork and combine a translation in X and a rotation around Y. Finally, the 6th mode is a translational mode in Z direction. A frequency response (SOL111) analysis is carried out with a unit acceleration excitation in each translational and rotational direction. The X, Y, Z displacements and angles of the mirror center of gravity are monitored giving 36 transfer functions of the mechanism. The responses that are the most detrimental to the pointing stability are the Y rotation (elevation) due to X -translation excitation, and Y rotation due to Y-rotation excitation (see Figure 9). Figure 9: Ball bearing pointing mechanism response on elevation axis rotation (Y-rotation) for a X - translational (left) and a Y -rotational (right) unit acceleration excitation Figure 9 shows that even though the mirror is perfectly balance d, micro -vibration at the base of a stiff pointing mechanism will be amplified by the structural mode and lead to rotation of the mirror around elevation axis. These perturbations cannot be compensated by drive electronics as they are taking place above the control frequency bandwidth. In addition, the prediction of the induced movement can be dif ficult, as the mode shape is playing a key role. In an optimistic approach, where the mirror is perfectly balanced and the first main mode located in a rather quiet micro -vibration frequency range where acceleration is 0.001 m/s2, the micro -vibration -induc ed motion around the elevation axis is o n the order of 1 µrad for X translation excitation. On top of this, the motion due to Y rotation excitation shall be added as well as all the noise coming from the mechanism ball bearings and the induced motion from unbalanced mass . In these conditions, it is difficult to reach the stability requirement of 0.1 µ rad or smaller. Passive Damping of Micro -vibration A simple way to reduce the micro -vibration effec ts on the mirror pointing stability is to introduce compliance in the system. The goal is to control and lower the rotational eigenfrequencies of the pointing mechanism below micro -vibration excitation frequencies from the spacecraft but still above control loop frequenc ies. A pointing mechanism with such low resonance frequency is passively damping the micro -vibration thanks to the natural transfer function decay observed above resonance frequency. Compliance is efficiently introduced in the mechanism using flexible pivots. A novel concept of a pointing mechanism has been developed to cover a full hemispherical pointing range . A fully functional, and motorized breadboard has been built to verify the functional and range requirement and is shown in Figure 10. The beauty of this concept is that it replaces the traditional coarse and fine pointing mechanism

Document

AMS_2020.pdf

512 used in high accuracy pointing mechanisms by one unique ultra- stable pointing mechanism with micro - vibration isolation capability, bringing mass, volume and cost down. Figure 10: Novel concept of pointing mechanism The pointing mirror is supported by two large angle flexible pivot s similar to the D3 design presented earlier and provides the elevation rotation around X. The ir stators are connected to a frame structure by two small angle flexible pivots (3rd Axis) to lower the Y rotation eigenmode of the elevation axis. The stator side of the small angle 3rd axis pivot s is then connected t o the azimuth fork . The azimuth fork is then connected to the azimuth ball bearing by a set of small angle flexible structures for decoupling the azimuth rotation axis from azimuth motorization mode, azimuth ball bearing, and spacecraft -born rotational micro-vibration . A flexible pivot is used between the elevation stepper and elevation axis to filter out the high frequency micro -vibration noise coming from the motor. The novel pointing mechanism dynamic behavior is studied using a simplified Nastran model shown in Figure 11.

Document

AMS_2020.pdf

513 Figure 11: Simplified Nastran model of the novel pointing mechanism The exact same structur al properties have been used for the novel concept using flexible pivot s as in the ball bearing version presented previously . The notable difference is that the axial, radial, bending and torsional stiffness of the flexible connection have been tuned to lower the first eigenmodes. Table 3 shows the eigenfrequencies of the first 12 modes together with their participating mass and inertia. Table 3: Novel concept pointing mechanism using flexible pivots eigenfrequencies and participating mass ID Freq TX TY TZ RX RY RZ Hz kg kg kg kg m2 kg m2 kg m2 1 0.2 7.51E -10 1.78E -10 8.90E -29 1.21E -16 1.52E -20 4.53E -02 2 0.4 1.90E -11 2.13E -27 2.41E -10 2.62E -28 3.53E -02 7.38E -22 3 4.3 2.76E -28 1.28E -09 6.58E -19 .272E -02 3.35E -27 3.73E -26 4 20.6 7.84E -25 1.22E+00 1.51E -16 6.98E -10 5.15E -29 5.45E -25 5 20.6 1.01E -15 2.92E -17 1.19E+00 3.55E -19 7.84E -23 5.34E -30 6 29.2 1.20E+00 3.56E -25 1.16E -15 2.19E -27 6.93E -12 2.24E -25 7 46.9 7.46E -11 1.96E -11 1.71E -25 1.64E -17 9.07E -21 2.82E -12 8 47.6 3.96E -12 2.01E -28 3.08E -11 4.42E -28 1.81E -11 7.78E -31 9 180.6 5.94E -20 5.47E+00 9.60E -12 2.09E -05 5.75E -28 2.96E -24 10 362.4 5.51E+00 9.92E -21 2.97E -10 5.12E -22 6.80E -07 1.35E -24 11 839.4 2.97E -17 1.26E -02 5.11E -07 1.34E -02 7.56E -24 1.03E -24 12 870.8 3.06E -10 9.74E -10 5.44E+00 1.08E -09 8.22E -17 7.64E -32 Modes number 1 and 3 are the same ones as those of Table 2, and are rotational modes around azimuth and elevation ax es. Mode number 2 is a new mode introduced by the 3rd axis which will greatly help to damp Y rotations of the mirror under perturbations coming from the spacecraft interface. Modes 4, 5 and 6 are translational modes of the elevation axis. Thanks to the low radial and axial stiffness of the large- angle flexible pivot, it is possible to lower these modes below the excitation frequencies of cryocooler s while keeping them well above typical controller frequency and avoid control issues.

Document

AMS_2020.pdf

514 Modes 7 and 8 are rotational modes of the elevation axis around the mirror Co G. Modes 9, 10 and 12 are translation modes of the azimuth fork and upper fra me combined . Benefits of decoupling the elevation axis from the rest of the structure can already be observed by looking at the participating mass. Indeed, the participating mass of these modes is close to the total mass minus the mass of the elevation axi s. A frequency response analysis (SOL111) is performed with unit acceleration excitations at the mechanism interface with the spacecraft. The transfer functions between mirror Co G motion and excitation are calculated and compared to the ones obtained wit h the ball bearings model . In all cases, significant ly lower responses are observed. To illustrate this, the two transfer functions shown in Figure 9 are reproduced in Figure 12 together with the ones obtain ed with the flexible pivot s concept. Figure 12: Flexible pivot s pointing mechanism response (in blue) and ball bearing s mechanism (in red) on elevation axis (Y -rotation) for X -translational (left) and Y -rotational (right) unit acceleration excitation The transfer functions show strong attenuation of the micro -vibration . Most of the pea ks are located in the 10 to 50 Hz bandwidth which is out side of typical micro -vibration spectrums . In the bandwidth between 50 and 200 Hz corresponding to intermediate micro -vibration excitation from cryo- cooler and reaction wheels of 0.001 m/s2, the maximal mirror rotation around Y is 6E-6 µrad at 47Hz , 6E-7 µrad at 50 Hz , and 3E-9 µrad at 200 Hz. Note that these values are for Ry excitation and are smaller for X excitation. The stability of a compliant pointing mechanism is thus between 3 and 8 orders of magnitude better than classical ball bearing pointing mechanism. To further improve pointing stability , dampers are added on the elevation, azimuth and 3rd axis. These dampers are acting in rotation but also in radial directions. The effect of these damper s decreases the local amplification at resonance peak. The elevation rotation due to X excitation can be reduce by a factor 5 on the mode at 29 Hz and by a factor of 10 for the mode at 49 Hz.

Document

AMS_2020.pdf

515 Figure 13: Flexible pivot pointing mechanism response with dampers (in light blue) and with ball bearing (in red) on elevation axis (Y -rotation) for a X -translational (left) and a Y -rotati onal (right) unit acceleration excitation Conclusion The development of a novel, highly customizable flexible pivot is opening new design possibilities. Thanks to a fully automatized optimization software, it is possible to design within a few hours a tailor-made flexible pivot that matches specific stiffness, range of motion, and lifetime requirements. In addition, it is possible to cover a wide range of requirement s, from low torsional to high torsional stiffness and from high to low torsion to radial stiffness. The application of flexible pivot s is thus numerous, from high- precision pointing mechanisms , to infinite- life oscillation scanner or to energy storage applications. By combining large- angle and small -angle flexible pivot s, micro -vibration excitation can be ruled out from the pointing accuracy budget. It is shown that it is possible to lower the micro -vibration impact on the pointing stability by 3 to 8 orders of magnitude between 50 to 200 Hz by the introduction of complianc e at strategic locations . In the frequency range above 200 Hz , which is the noisiest bandwidth of reaction wheels, such a novel concept is bringing the level of perturbation angle below 5E -10 rad/m/s2. The addition of dampers on the three axes can furthe r improve the pointing stability by a factor of 5 to 10 in comparison to the pointing mechanism using flexible pivot s alone. It is thus possible to replace the combination of coarse and fine pointing mechanisms by a single mechanism achieving ultra- stable pointing stability. The total mass of the system can be drastically reduced, as well as the total volume. Finally, the novel ultra- stable pointing mechanism using flexible pivot s is very cost effective in comparison to solutions available today on the mark et. The highly customizable flexible pivot design family is an enabling technology for many applications such as ultra-high precision pointing mechanisms . References 1. L. Blecha, M. Humphries, Y. Puyol, WO 2017077469A1 2. Schlüter, M. &. (2010). The or acle penalty method. In Journal of Global Optimization, 47(2) (pp. 293325). 3. Schlüter, M. E. (2009). Extended ant colony optimization for non- convex mixed integer nonlinear programming. In Computers & Operations Research, 36(7) (pp. 2217- 2229). 4. Schlüter, M. G. (2012). A numerical study of MIDACO on 100 MINLP benchmarks. In Optimization, 61(7) (pp. 873- 900). 5. Puyol, Y. (2019). Innovation in Large angle flexible pivot Design & Material Accelerated Fatigue Screening Tests Results. In ESMATS 2019 .

Document

AMS_2020.pdf

517 Compliant Mechanisms M ade by Additive Manufacturing Lionel Kiener*, Hervé Saudan* , Florent Cosandier *, Gérald Perruchoud *, Vaclav Pejchal *, Sébastien Lani* and Antoine Verhaeghe* Abstract Several compliant mechanisms have been completely redesigned for Additive Manufacturing (AM) and have allowed CSEM to develop an innovative concept. In addition to the new geometric possibilities offered by AM, the need for machining and assembly after pri nting are drastically reduced. Support structures under flexure blades are thus minimised and the overall process becomes more streamlined. Moreover, this idea allows us to easily design and produce monolithic cross blade flexure pivots with interlocked fl exible blades. Thanks to this concept, CSEM is now developing and testing new architectures of Compliant Mechanisms based on Additive Manufacturing (COMAM) for the European Space Agency (ESA). The development methodology, the AM process and post -process and the testing approach are detailed in this paper. Figure 1. Compliant Mechanism built by Additive Manufacturing. Introduction Mechanisms with friction present significant drawbacks with the need of lubrication, debris generation, backlash and stick slip. In cryogenic and space environments, suitable lubricants are very limited when not prohibited. Wear generation can pollute optics, obstruct a smooth motion and c an even lead to early failures. To overcome these important limitations , Compliant Mechanisms (CM) are usually proposed. They can achieve macroscopic linear and rotary motion without friction, wear, backlash, and with extremely high fatigue performance thanks to the elastic deformation of flexible structures. They are used in harsh environment s such as vacuum, cryogenic and space combining high- precision and a long lifetime capabilities . * Centre Suisse d'Electronique et de Microtechnique (CSEM), Neuchâtel, Switzerland ; lki@csem.ch Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 2020

Document

AMS_2020.pdf

518 To date, the extreme complexity of compliant mechanisms has required highly sophisticated and expensive manufacturing methods, the gold standard being the Wire Electro -Discharge Machining (WEDM) from a bulk material block with consecutive large material losses and very long and delicate machining procedures. Moreover, the assembly has actually to be realized with many precautions to ensure a very precise positioning between all parts. Today, this paradigm is questioned by the possibilities offered by AM technologies, notably the metallic powder bed processes such as the Selective Laser Melting (SLM). While the largest part of the research presently repor ted is focused on developing and optimizing designs of what could be described as “structural or massive parts”, little work has been published up to now to determine the limits related to the manufacturing of thin, flexible structures used in compliant mechanisms [1]. After more than 30 years of successful developments using compliant mechanisms produced by conventional manufacturing methods, CSEM demonstrated in 2016 the feasibility of high performances compliant structures made by AM [ 2]. Over the last few years, CSEM has acquired an expertise in the computerized optimization of such mechanisms for AM and has proceeded further by inventing a totally new design concept: interlocked lattice flexures. This new type of compliant structure geometry and arrangement is such that the flexure elements cross, but never touch each other, even when deformed. This new architecture – made only possible by AM technologies – creates the opportunity to develop completely new flexure topologies but also to improve existing ones, as demonstrated with the example of a redesigned C -flex type pivot (patent US 3073584) illustrated in Figure 2. Figure 2. Example of the redesign of a C -flex type pivot with interlocked flexure blades. Compliant Mechanisms Heritage CSEM is active in the design and development of very high performance flexural elements and mechanisms for more than 30 years. Notable examples for space applications are the HAFHA flexural pivot and the Corner Cube Mechanism which is currently operated in the IASI instrument on board MetOp satellites, to date wi th more than 1 billion cycles (linear stroke of ±15 mm) achieved in 14 years. Two flight models of Corner Cube Mechanism has also been delivered last year for the infrared sounder onboard two Meteosat Third Generation satellites. Another example is the CLU PI linear Focus Mechanism for ExoMars rover. Other mechanisms (e.g. , Slit Mask, tip -tilt and chopper) have been developed and produced for ground based telescopes as well as for the airborne SOFIA telescope.

Document

AMS_2020.pdf

519 In the same philosophy, the elaboration of new mechanisms made by additive manufacturing has been investigated at CSEM over several years targeting the general goals of assessing the benefits and weaknesses of the AM fabrication process for compliant mechanisms and getting a sufficient level of expertis e on AM produced compliant mechanisms in perspective of future projects. Design Methodology for AM -Based Compliant Mechanisms The methodology to develop C ompliant Mechanisms built by AM is not straigtforward since the software tools do not always have a sufficient maturity for these kind of systems. The development workflow has been developped by CSEM to best utilize the strengths of all design, AM process and post -process simulation soft ware. It consists of multiple steps as described herein . First, the preliminary design is performed with the definition of the global compliant architecture and a preliminary sizing. Then, the design is refined in two parallel processes: topology optimizat ion of the rigid structure and shape optimization of the flexures. Finally, complete Finite Element Modelling simulations are performed to verify the compliance to the requirements. The principal steps of the design flow that have been elaborated by CSEM to successfully achieve the development of a compliant mechanism based on AM are presented hereafter and illustrated by the example of the Compliant Rotation Reduction Mechanism (CRRM) shown in Figure 3. Figure 3. Compliant Rotation Reduction Mechanism (CRRM). The development of this CRRM is made for the European Space Agency (ESA) for a research project. Specifications The principal specification for the CRRM is that the mechanism shall be totally frictionless. In terms of performance, the input angle shall be ±10° while the output angle shall be ±1°, meaning that the reduction ratio of the mechanism shall be 1:10. The input and output are also inverted, for a 10° clockwise rotation, the 1° output i s counter clockwise. The repeatability of the system implies that the parasitic motion at output shall be smaller than 10 µm in the lateral and axial directions and that the parasitic tilt shall be smaller than 1/100°. Its dimensions shall be 120 mm x 50 mm and its mass shall be a maximum of 0.4 kg. For environmental performances, the mechanism shall withstand launch sinusoidal vibrations of 24 g, random vibrations of 18.4 gRMS and shocks of 1000 g.

Document

AMS_2020.pdf

520 Preliminary design and trade- off The preliminary design activity of an AM- based compliant mechanism can be divided into two phases. The first one consists in conventional pre- design activities. The flexure topologies and the overall physical architecture forming the basis of the design are defined, involvi ng the analytical pre- sizing of various alternatives. A pre- design example of the CRRM is given in Figure 4. Figure 4. Architecture and pre- design of the CRRM. Design for Additive Manufacturing This pre-design is then considered under the perspective of the manufacturing process, i.e. , Selective Laser Melting (SLM). The following aspects are choosen during this phase: • optimum build- up orientation, • identification of the critical geometries , • geometry of interfaces; fixation areas, positionning features, reference surfaces, • AM process strategy (support material and its future separation from the part) . This is performed by taking into account support structure minimization in critical locations – where postAM machining could be difficult if not impossible, post -process strategy (thermal treatment before/after removal) and separation from the build plate. These activities are realized in accordance with the general design rules for AM and the specific rules for compliant structures which have been developed at CSEM. The manufacturability of the design should then be assessed. This is done thanks to SLM process simulation software. A post -processing sequence, including thermal & mechanical post -process es and a verification strategy is defined in accordance with the specific requirements for compliant structures, such as temporary fixation of mobile stages and the addition of features for metrology. Detailed design The detailed design comprises two main phases: • Topology optimization of the rigid structure, • Shape o ptimization of the compliant structure, i.e. , the flexure blades. Rigid structures optimization A topology optimization of the rigid structure is performed on the initial design in order to i mprove its mechanical characteristics, especially the overall rigidity, together with a mass reduction. The workflow is the following: 1. Definition of the design and non- design spaces, where the design space is the part of the item where the optimization solver will be active. The non- design spaces are mainly the interfaces and other peculiar locations which need to be conserved as defined in the preliminary design. See Figure 5

Document

AMS_2020.pdf

521 2. The boundary conditions and the load cases are defined. 3. The optimization parameters are defined. 4. The results are interpreted. 5. A shape smoothing and/or rebuild is performed at the end as illustrated in Figure 6 6. A final finite element analysis with the new shape is performed to ensure fulfilling the requirements . Figure 5. Definition of the design spaces for the CRRM . Figure 6. Result of the topological optimization ( left); design example after smoothing ( right). Flexure blades optimization The compliant structure shall be optimized separately to ensure an optimum solution with regard to performances, but also to ease as much as possible the manufacturing and the post -treatments, mainly the removal from the build plate. The need to include suppo rt structures while producing thin flexure blades by AM is a critical aspect that must be taken into account while designing CM. The support structure is minimized and the attachment points of the support structure to the flexure are weakened in order to m ake its removal easier. The separation is performed when the part is cut off from the build plate. This concept has been successfully tested with several designs . Top viewOutput (in pink ) Fixed interface (in grey ) Bottom viewInput (in light red )

Document

AMS_2020.pdf

522 As no single solution allows for simultaneously optimizing the rigid and the flexible part of the mechanism [3, 4], a dedicated procedure is devoted to this task. Lattice flexure blades While looking for the most appropriate design for flexure blades, CSEM innovated with a lattice structure (patent pending) having the main advantages of: • Lowering the bending stiffness while maintaining a sufficient thickness for manufacturing, • Avoiding internal support structure thanks to the overhang angle, • Ability to be interlocked to form a pivot. We start by defining a unitary lattice cell from which the whole blade pattern will be generated applying symmetry operations. Then, this unitary cell is geometrically parametrized. Next, a large number of different cells are generated using a Monte Carlo method. Some rules must be respected regarding the manufacturing and integrity of the structure. Therefore, only the designs that are compliant to those rules are considered. For these remaining solutions, an objective function is defined based on differen t mechanical parameters with dedicated weighting factors. Example of such parameters are transverse stiffness and stresses. Another criterion to be assessed is the constancy of the section area along the longitudinal axis of the leaf spring. The goal is to select a lattice that has a cross -sectional surface as constant as possible in order to avoid having a polygonal effect, to maintain a constant curvature of the leaf spring and to mimic at best the behaviour of a plain leaf spring. Finally, one of the r emaining designs is selected as candidate for the final, detailed design, as shown in Figure 7. Figure 7. Stress distribution for one particular design (left); optimal lattice leaf spring pattern (right). Interlocked lattice flexible structures Thanks to these optimized lattice structures as well as the opportunities given by AM, interlocked lattices flexures as illustrated in Figure 8 can be proposed. This architecture forms a monolithic rotational pivot with a high axial stiffness and which can be additively build with very little support structure.

Document

AMS_2020.pdf

523 Figure 8. Monolithic r otation pivot composed of two latticework blades (patent pending) . Based on these interlocked lattice flexures, the CRRM has been designed and the first prototypes of the CRRM have been successfully produced by AM- SLM. Final simulation results The simulation resuts after optimization show that the Statement of Work requirements are globally fulfilled, as presented in Table 1. Table 1: Comparison of requirements with simulation results Main requirements Statement of Work Simulation results Input angle ±10° ±10° Output angle ±1° ±1° Diameter 100 mm 120 mm Length 50 mm 40 mm Center shifts < 10 µm < 350 µrad < 2 µm 3 µrad First eigen mode > 100 Hz with blocked IF 740 Hz (400 Hz before optim.) Input torque to be minimized 0.24 Nm Lifetime min. 100 ,000 cycles 1 mio (goal) Infinite lifetime at ±10° Material Vacuum compatible, -60° to +80°C Stainless steel 17- 4PH

Document

AMS_2020.pdf

524 Manufacturing Manufacturing assessment The additive manufacturing assessment has been made with the help of the Amphyon software tool, which simulat es the Selective Laser Melting (SLM) process to give information about the internal stresses generated during layer manufacturing. These stresses c ould be responsible for macroscopic deformations of the parts as shown in Figure 9. Figure 9. Manufacturing layer by layer thermo- mechanical simulation. Based on these simulations, this tool generates a pre- deformed 3D geometry of the part to overcome these deformations with the aim of having a geometry that conforms to the nominal ly designed shape. Based on preliminary tests with thin and flexible structures, the residual deformations were in the range of 0.1 mm. Material, process and post -process t esting The preliminary material, process and post -process test results have already been presented during ESMATS 2017 [3]. During the current COMAM project, these results have been consolidated with new tests such as residual stresses, dissolved gases, tensile, hardness, roughness, general corrosion, stress corrosion crac king and fatigue. In parallel, the microstructure was verified as well. These tests are performed on representative samples which have been additively manufactured in a high- strength stainless steel 17- 4PH. They have seen the same post -processing treatment s as foreseen for the final mechanism (i.e., Hot Isostatic Pressing (HIP), solution annealing and age hardening). In complement, the entire COMAM mechanism will be tested following the classical space approach with performance, vibrations, shock and thermal cycl ing. The first results are presented in the next chapter Mechanisms testing . Tensile test results Ten tensile samples machined out of AM -built cylinders were characterised. At room temperatures, measured values of Yield strength (R p0.2) and Ultimate tensile strength (UTS) were very similar for all tested samples and varied from 1280 to 1330 MPa and 1380 to 1450 MPa for yield strength and UTS, respectively (Figure 10). For comparison, typical values for extruded forms are Rp0.2: 1070 MPa and UTS 1170 MPa. The yield strength was slightly higher at 1410 MPa and 1440 MPa for samples tested at - 40°C while UTS remained relatively unchanged. Measured Young’s modulus E is between 190- 210 GPa.

Document

AMS_2020.pdf

525 Elongation at failure exhibited the highest degr ee of variation from 1.2 to 6 %. Fractography revealed the presence of lack -of-fusion defects in the specimen with the lowest elongation (1.2%). For the rest of the samples tested at room temperature, necking occurred outside the measured gauge length, whi ch contributed to the overall spread in measured elongations. At low temperature ( -40°C ), the ductility remains relatively high reaching near ly 7%. Figure 10. Left: s tress -strain curves of six tensile tests performed at room temperature; right: hardness measurement results. Hardness test results Micro -hardness was measured on both ends of tensile samples after machining from cylinders. HV0.3 results lie within 450 and 500 which is a spread in values typical for micro -hardness measurements (ca. 10%). HV0.3 between 450 and 500 corresponds to approximately 48 HRC which is near the upper end of expected hardness values of 17- 4 PH for this thermal condition. Roughness test results The surface quality has been measured with a surface roughness tester on the fatigue test samples. No mechanical process has been performed on the surface. The mean Ra value is 8 µm (±1.5 µm) and is independent of the direction of printing and of the thermal treatments performed after printing. Compared to surfaces obtained by machining, this value could be seen as much higher but the roughness is only an indicative value. The fatigue test results are much more important with regard to the behavior of the compliant mechanism. Fatigue test results The fatigue behavior of this material has already been defined during a previous activity at CSEM with an alternate bending fatigue test bench. Additional fatigue tests have been carried out to consolidate the results, including the lattice flexure blades. The results indicate that the values of these AM-flexure blades are comparable to the results previously obtained by CSEM [2].

Document

AMS_2020.pdf

526 Mechanisms t esting Cleanliness assessment The whole build plate was thoroughly cleaned after SLM in order to remove unfused metal particles and other potential contaminants before the HIP treatment. During the cleaning process performed in ultrasonic (US) bath, cavitation was visibly very homogenous, which indicates optimum US exposure . A significant amount of metallic particles was collected. In total 0.55 g for a total build mass of 1251 kg, representing 0.044%. This process was repeated after HIP to see if more unfused particles could be removed. Here, only few particles have been collected for a total mass of 0.03 gram. They can be classified in four catergories: raw powder, round dark particles, flakes (probably contaminants during HIP process) and a few bigger particles which should be partially melted powder agregates , as shown in Figure 11. Figure 11. Collected particles after HIP. Metrology A combination of 2D optical metrology and 3D laser scan has been performed in order to evaluate the global build plate deformations which are mainly due to the stress generated during the SLM and thermal post-process es (HIP, solution annealing and age hardening). The metrology has been performed first after SLM, then after HIP and finaly after SA -AH (still to be done at the time of writing). The comparisons are done first with the 3D CAD model , then between the step before and the actual state. The first re sults are shown in Figure 12. More work is ongoing to assess the impact of these deformation on performance. Figure 12. Left: global deformation of the build plate after SLM. Right: overview of the deformation for one mechanism. Scale in mm.

Document

AMS_2020.pdf

527 Next Steps The next step of the COMAM project is the final metrology measurements following the thermal treatments and the machining. This will be followed by performance measurements, vibration, shock and thermal cycling testing. The performance measurements will be compared with the simulation results to validate the design and ensure the compliance with the requirements. Conclusions We have described CSEM’ s methodology developed to design, optimize and verify the development of an innovative compliant mechanism made by additive manufacturing. We have sur passed challenges and pushed this technology forward to implement innovative solutions. The ESA COMAM project is ongoing. The next steps are the manufacturing of two Elegant Breadboard Models followed by the test campoaign ; performance, vibration, shocks, thermal cycles and lifetime. In parallel, the testing of the characterization samples is in progress. CSEM continues to work on the ultimate goal to have a global tool for the optimization of compliant mechanisms. In parallel, functionalization of AM parts has been demonstrated and further projects will bring new examples of adding electric al, thermal and optical features. References 1. Merriam, Ezekiel G., Stiffness Reduction Strategies for Additively Manufactured Compliant Mechanisms, All Theses and Dissertations. Paper 5873, 2016. 2. Saudan, H., Vaideeswaran, K., Kiener, L. & Dadras, M. (2016 ). Additive Manufactured Metallic Flexible Structures, a focus on Manufacturing Strategies, Material Analysis and Fatigue Verification. In Proc. European Conference on Spacecraft Structures , Materials and Environmental Testing, Toulouse, France, 27-30 Sept. 2016. 3. Saudan, H., Kiener, L., Perruchoud, G., Vaideeswaran, K. & Dadras, M. (2017). Additively manufactured and topologically optimized compliant mechanisms: technological assessment approach, latest achievements and current work in progress. In Proc. 17th European Space Mechanisms & Tribology Symposium, Hatfield, United Kingdom, 20- 22 Sept. 2017. 4. Saudan, H., Kiener, L., Perruchoud, G., Vaideeswaran, K., Dadras, M. & Cochet, F. (2018). Compliant mechanisms and space grade product redesign based on Additive Manufacturing. In Proc. SPIE 10706, Advances in Optical and Mechanical Technologies for Telesc opes and Instrumentation III, 107062T .

Document

AMS_2020.pdf

529 Flexible Waveguides for RF Transmission across PSP HGA Rotary Actuator Deva Ponnusamy*, Weilun Cheng* , Ted Hartka* , Devin Hahne *, Calvin Kee *, Mike Marley * and David Napolillo * Abstract The High G ain Antenna ( HGA ) on the Parker Solar Probe ( PSP) spacecraft was mounted on a s ingle axis rotary actuator with a range of motion of ±45 degree. During the early phase of the program, a trade was performed to select the appropriate technique to manage the Radio -Frequency (RF) trans mission across the rotary joint . The Flexible WaveG uide (FWG) option seemed attractive, due to its low mass and good RF characteristics . However , the performance of these waveguides across rotary joints under repeated articulation was not well understood. So a development program was conducted and flight -like FWG assemblies were s ubjected to various test s. The successful performance of the waveguide resulted in the selection of the FWG for the PSP HGA assembly. As the flight design progressed, additiona l structural and thermal analyses were performed on the HGA assembly to evaluate the use of these waveguides . The flight batch of waveguides were subjected to a comprehensive batch qualification and screening program. Special tools and procedures were developed for installation of the flight waveguides on to the HGA assembly . The PSP spacecraft was launched in August 2018 and the FWG s have been performing well as planned. The paper discusses the development, qualification and flight activities and the key lessons learned along the way. Introduction The PSP spacecr aft is equipped with a Ka-band Cassegrain- type High Gain Antenna as shown in Figure 1. The main reflector is a 0.6-meter diameter composite dish with a notched area on the edge for increased dynamic clearance to the spacecraft in stowed configur ation Figure 1. PSP HGA Assembly * Johns Hopkins University Applied Physics Laboratory, Laurel, MD Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 2020

Document

AMS_2020.pdf

530 The reflector assembly i s mounted on a single axis rotary actuator with a range of motion of ±45 degree. This system required two RF channels for receive and transmit across this rotary joint , and accordingly the HGA assembly was designed to accommodate two FWG s connect ing the feed assembly to the RF module inside the spacecraft through the bulkhead waveguide. The use of the FWG was explored and validated by a development program which consisted of detailed analysis and extensive testing . The structural, thermal and life cy cle characteristics of the waveguide were evaluated with specially developed tests and techniques . This program also experienced a process issue during the flight batch qualification. In coordination with the manufacturer, Custom Microwave, the root cause was quickly identified and an improved process was employed, resulting in a robust component. Trade Study The PSP HGA actuator had a relatively smal l range of motion and a low life cycle requirement of 110 cycles but was required to fit within a tight mass budget and envelope. A trade study was performed to identify the most suitable RF transmission technique. The RF requirement precluded the use of a flexible coaxial cable. The options considered for this trade were a n RF Rotary Joint and an FWG. The waveguide c onsidered for this application wa s electroformed Ni -Co waveguide. This ty pe of flexible waveguide is highly elastic and is capable of repeat ed cycling in bending without permanent deformation. The PSP RF system required a WR-34 waveguide and the flexible length was chosen in consul tation with the vendor to minimize bending stresses while limiting launch loads due to vibration environment. The RF rotary joint is an electromechanical device consisting of a rotor assembly mounted to the stator through duplex bearing pairs. The part con sidered for this trade was a standard WR -34 joint with waveguide interfaces at input and output. These devices have good RF performance and spaceflight heritage. Both options were evaluated for mass, volume, alignment requirement and heritage for the prop osed usage. − The mass of each w aveguide was 40 grams . The mass of each RF rotary joint was 300 grams and the two waveguides were mounted to a bracket of mass 250 grams . The tot al mass of the RF joint system wa s 850 grams compared to 80 grams for the FWG s. − The FWG option required a much smaller envelope than the rotary joint assemblies . Accommodation of two of the rotary joints along with the routing of the waveguides required a larger envelope. This envelope was particularly significant for the PSP mission since the volume available for HGA accommodation was very limited. One of the unique features of the PSP spacecraft is a Thermal Protection Shield (TPS), which was a sandwich panel made up a ca rbon- carbon composite facesheets and a carbon foam core. During close flybys of the Sun, the spacecraft is oriented with the TPS facing the Sun so as to protect all spacecraft components within the umbra of the TPS. This need to fit within the umbra of the TPS imposed severe restrictions on t he envelope. − RF rotary joint s consist of preloaded duplex bearing pairs and require very tight alignment with the actuator axis. Any misalignment due to installation or on- orbit thermal deformations would impact the RF performance of the rotary joint as well as the actuator mechanical performance. The FWGs can tolerate higher misalignments relative to the rotary axis. − The cold operational temper ature of the RF joint is driven by the bearing lubrication which is typically about - 40°C for wet lubricated sys tems . The flexible waveguide can be operated at much lower temperatures. − RF rotary joints have been used successfully for thousands of cycles on multiple space missions. The FWG considered for this application has extensive flight heritage but its primary function was to accommodate misalignments. There are a few instances of FWGs used on one-time deployment mechanisms . However, there is very li mited information on the use of flexible waveguides for cyclic operations in spac e over the life of the mission.

Document

AMS_2020.pdf

531 The FWG seemed better on every criterion except heritage. So a development testing program was implemented to evaluate the FWG for PSP requirements . The FWG performed successfully in these tests and was chosen for the HGA assembly . The development test details are discussed later in this paper. Flexible Waveguide The flexible waveguide considered for this application is made by Custom Microwave. Inc. of Longmont, C O. The flexible part of the waveguide is Ni-Co alloy and is manufactured by electroforming. This waveguide was built to length specified by JHUAPL. The interior of the waveguide is coated with silver to improve RF performance. The development test article had copper flanges that were attached to the waveguide by soldering . Operational high temperature of the waveguide was limited to +180°C due to this soldered joint . The exterior of the waveguide was nickel plated . During the development phase, the qualification temperature range was - 125°C to +100°C and t he development article successfully survived these temperatures . However, by the time the flight articles were procured the qualification temperature limit had changed to - 105°C to +170°C. At this time the manufacturer had improved processes available, but the program decided to stay with the qualified version of the waveguide , since the revised temperature was still within specifications . The first batch of flight waveguides were identical to the development article and was tested to the revised temperatur e limits. During thermal cycling some discoloration was observed at the soldered joint, along with significant degradation in RF performance. Investigation by the manufacturer determined that this was due to a processing defect that resulted in incomplete removal of the flux from the solder , and this was also confirmed by radiography . This problem was further exacerbated by the higher revised qualification temperature. The manufacturer proposed waveguides with electroformed flange joints, which were considered superior to the soldered joint. After a thorough review of all options it was decided to change the waveguide configuration to the electroformed flange joint as proposed by the ma nufacturer . The program also decided to switch the exterior finish to black paint inst ead of nickel plating to lower the temperature of the middle section during operation. Figure 2. WR -34 Flexible Waveguide The flight waveguide was made up of electroformed Ni-Co flexible section with the copper flanges attached by the electrofor ming process. T he interior surfaces were silver plated and the exterior surfaces were painted with BR -127 black paint . Following a successful batch qualification and acc eptance program, these waveguides were installed to flight HGA assembly. The development and flight waveguides are shown in Figure 2. Waveguide Analyses and Test s The FWG was subjected t o a comprehensive evaluation program that included testing in three phases, development test, flight batch qualification test and acceptance tests. These test s consisted of environmental tests and mechanical functional tests. The condition of the FWG was monitored by RF

Document

AMS_2020.pdf

532 performance tests, nondestructive examination and visual inspection. Following is a brief description of the various inspection and test techniques employed in the test program. Resistance Torque and Flex Life Cycle Resistance torque was measured over the range of motion on a torque test rig as shown in Figure 3. The maximum resistance torque was less than 30 mN -m and did not have an influ ence on the torque margin. The resistance torque was also considered as a good m easure of the uniformity of the waveguide thicknesses and was viewed as an indicator of consistency and process control. Figure 3. Resistance Torque and Life Cycle Test The stresses induced due to the ± 45-degree motion wer e well below the elastic limit, yet life tests were performed to ensure that cyclic loading does not cause any damage or performance degradation. The cycling test was performed at ambien t, hot and cold temperatures, in ambient GN 2 atmosphere (Figure 4) . The test article was flexed insi de a thermal chamber by a drive system made up of a motor and a torque sensor . The flight batch qualification unit was subjected to 330 cycles, equally divided between ambient, +170°C and -105°C. The resistance torque was mo nitored to observe any trend. Structural Evaluation The key structural feature of the FWG is the thin, corrugated Ni -Co wall of the waveguide that allows it to flex. The waveguide length was chosen to minimize the elastic stresses due to the cyclic motion . The long unsupported length also meant lower resonance frequencies and higher load due to launch environments. One particular area of interest was the natural frequency of the waveguide in its l aunch configuration. The intent was to not have the waveguide resonances couple with the spacecraft frequencies. In the launch configuration, the high gain antenna is stowed at 45°, which put the waveguides in a gentle curvature (Figure 5).

Document

AMS_2020.pdf

533 Figure 4. Life Cycle Test at Temperature Extremes During the development phase, a set of six waveguides with soldered flanges were subjected to low -level sine sweep from 5 Hz to 500 Hz, to determine the natural frequencies in the off -axial directions. The waveguides were mounted on test blocks that flexed the wavegu ides by 45 degree s flange- to-flange, as in launch configuration. A miniature, single -axis teardrop accelerometer was attached to the middle of each waveguide, aligned in the direction of the test axis. There was a concern that the accelerometer attachment might by itself influence the resonance of the waveguide. To address this, the test was repeated on one of the waveguides with a non -contact laser vibrometer. There was no difference in the resonance values or amplification measured by the accelerometer and the vibrometer. All subsequent tests only used an accelerometer. The natural frequencies in the narrow section direction ranged from 51 Hz to 60 Hz, and the wide section direction ranged from 60 Hz to 65 Hz. Because of the curvature of the waveguide in the launch configuration, the waveguide is somewhat preloaded and therefore stiffer in the wide direction, which led to a higher natural frequency when compared to the narrow direction. The variation in natural frequencies between waveguides w as attributed to minute variations in wall thickness due to the electroforming process control . Regardless, the minimum resonant frequency was 51 Hz, which was well separated from the spacecraft frequencies, and would pose no problems with over -amplification of its acceleration responses during launch vibration. Figure 5. Vibration Test in 45 Degree Bend Configuration

Document

AMS_2020.pdf

534 Sine vibration was performed to qualify the strength of the FWG for its expec ted flight levels, 15 G in the spacecraft lateral axes and 20 G in spacecraft thrust axis. Any resonance response was limited to 36 G according to its mass -acceleration limit load. The strength verification was accomplished by applying a sinusoidal vibration at the qualification level in the frequency range less t han resonance so that the acceleration load approximated a static loading. All waveguides passed this test and demonstrated the capability to handle expected launch loads. The maximum response recorded was 37 G after notching (Figure 6). Random vibration t esting was performed on the FWG s to the protoflight levels . The development and flight batch qualification waveguides were exposed to a random vibration input of 6.3 Grms in the wide section direction, and the max imum overall response recorded was 16.8 Gr ms (Figure 6) . At 3σ, the acceleration response would be 50.4 Grms , indicating that the waveguide could withstand loads higher than the design limit of 36 g. In spite of the large responses seen in the tests, the w aveguide is relatively light (40 grams), and therefore the load supported in its own structure during vibration is less than 22 N . Figure 6. Vibration Test Responses Another area of interest was to assess the tolerance of the FWG to axial loads, either due to installation misalignments or on- orbit thermal mismatch. The maximum estimated extension with margin, due t o all possible sources was about 0.7 mm. One of the development model waveguides was subjected to 300 cycles of extension on a Universal Testing Machine and the load was monitored. For this extension, the waveguide was within the elastic limit and the load at maximum extension remained stable at 1.42 N. The waveguide was also subj ected to compression, the waveguide flexed slightly to accommodate this with very lit tle change in load. Thermal Concern A few thermal concerns with the us e of the flexible waveguide were identified during the development phase of this program. They were further explore d through analysis and test, resulting in design changes to both the HGA and the FWG . One of the key concerns was that the geometry of the FWG could cause the middl e section of the waveguide to become very hot. The waveguide had a relatively small cross- section and a large distance between the middle section and the flanges. The flexible part of the waveguide was very thin, and the corrugated path of this section created an effect ive conductive length that was several times the actual length of the waveguide assembly . Given that thermal resistance increases with length and decreases with area, the middle of the waveguide was thus strongly thermally isolated, in spite of being made from Ni -Co metal alloy . Even if flange temperatures were to be controlled, there still could be a very high temperature in the cen ter of the waveguide due to power dissipation during RF transmission. Adding to the concern with the middle temperature of the waveguide was the fact that this portion of the w aveguide could be exposed

Document

AMS_2020.pdf

535 to sunlight , depending on HGA position. A t certain times in the mission the sun exposure could occur at 0.7 AU, which would effectively double the solar flux on the waveguide. This was especially concerning given the optical properties of the nickel -plated exterior surface of the waveguide. Metals inherently have low emissivity, and do not radiate he at easily to space. M etals can get very hot in the sun if not properly sunk to another boundary temperature. The typical values for absorptivity -to-emissivity ratio for nickel from various sources was around 10 (0.4 / 0.04). This optical property, coupled with the high solar flux an d relatively high r esistance to the flange, would result in high temperatures in the center of the waveguide. The other key thermal issue with the waveguide was the temperature of the flanges. The HGA itself was allowed to run relatively hot with a desi gn limit of 160°C and this caused waveguide flange temperatures as high at 160°C . Thermal Analysis Analysis of the flex waveguide in Thermal Desktop software showed very high temperatures for the middle portions of the waveguide due to high thermal resistance between the middle of the waveguide and the flanges. A detailed model of the waveguide was created to estimate the temperature of the middle section with higher fidelity (Figure 7) . This model showed temperatures around 400°C depending on the diss ipation estimates. The temperature at the middle of the waveguide for the given flange boundary conditions and a high heat dissipation of 1.7 W was 416°C . Despite the high temperature at the middle, the model still showed the flange areas at reasonable tem peratures. The high temperature was not of particular concern from the waveguide material point of view, but could cause severe degradation or possible loss of the silver coating at 425°C. Hence it was decided to perform a test to verify the high temperature phenomenon. Figure 7. RF Induced Waveguide Temperature - Prediction Thermal Test A test was developed using a development waveguide in a vacuum chamber. This test used a Direct Current ( DC) voltage to apply dissipation across the waveguide. During different configurations of the test , the temperature was measured with thermocouples and with an Infrared ( IR) camera. Both methods were prone to high errors, the thermocouples at high temperature, and the IR camera on a low emissivity surface. However , the test was more to gain a general understanding of the issue , not necessarily to completely quantify the temperatures. The setup for the test in a thermal vacuum chamber is shown in Figure 8. In t his setup a thermocouple (T C) was placed in the middle of the FWG to get a direct measurement of the temperature. In another set up this TC was removed and photos were taken with an IR camera. The red and black wires attached to the flanges of the waveguide were connected to a power supply and used to run current through the waveguide and simulate the dissipation from an RF signal.

Document

AMS_2020.pdf

536 The test showed results that were somewhat consist ent with the analysis. Temperatures in the testing were not quite as high, however there was a concern that TC adhesion to the WG could have been affected at the high test temperature and resu lts might not have been completely accurate. The highest observed temperature of the waveguide TC during testing was 276°C. Temperature measurements were also done with the IR camera and the IR image is shown in Figure 8. This clearly showed the relatively cold temperature at the flanges and increase in temperature towards the middle. The testing was considered successful since it confirmed high temperatures predicted by analysis and also showed that the predictions were conservative relative to the test. The test also demonstrated that flange temperatures would stay close to their interface temper atures and would not overheat due to the dissipation in the waveguide. RF testing after the thermal vacuum test showed no degradation to the waveguide characteristics. Despite the fact that the test showed generally positive results, concern remained that additional heat on the waveguide from the sun could still result in excessive temperatures. For this reason, sun shields and a sun visor were added to the HGA assembly to prevent direct exposure of the waveguides to the sun , without interferin g with the waveguide or the swept volume. Another takeaway from the test was the significance of the RF dissipation and the desire to run RF power through the HGA assembly during the HGA level thermal balance testing. This testing was completed but is beyond the scope of this discussion. Figure 8. RF Induced Waveguide Heating - Measurement The exterior finish on the flight waveguides was changed from Ni plating to black paint and this further reduced the temperature of the middle section. The thermal cycling limits for the flight waveguides were established based on the flange temperatures. The waveguides were supported in the free- state and RF performance evaluated pre - and post - thermal cycling. As explained earlier, the upper temperature of the development model was underestimated. But this was revised prior to the flight article batch tests as the thermal design matured. Also due to the criticality of the thermal environment and the issues experienced with soldered waveguides, the flight batch quali fication article was subjected to a conservative 200 thermal cycles. The t hermal cycling limits of the various models were as follows: Development tests: 6 cycles, - 125°C to +100 °C Qualification tests : 200 cycles, - 105°C to +170 °C Acceptance test: 7 cycles, -105°C to +170 °C Waveguide Status and Performance Monitoring The condition of the waveguide was monitored at every stage of the test program to detect any degradation. This was done by visual inspection of exterior , inspection of interior with borescope, CT -scan, and RF tests.

Document

AMS_2020.pdf

537 Visual Inspection: All surfaces were visually inspected to keep track of any surface features or defects (Figure 9) . The exterior surfaces were inspected under 10X and 80X magnification. The condition of the silver plating on the interior had a significant influence on the RF performance and so the inside surfaces were inspected with a miniature borescope with a 90° aperture (Figure 10) . The borescope was mounted to a linear stage and the waveguide was supported on a stationary platform that allowed fine alignment of the waveguide with the bore scope. With this setup, the coating could be inspected without the risk of contact or damage. The wav eguides were screened by the manufacturer prior to delivery and no significant defects were noticed on the exterior or the interior. Any minor blemishes observed were photographed and tracked over the course of the test program. Figure 9. Visual Inspection of Exterior Surfaces Figure 10. Inspection of Silver Plating Radiography : The flexible parts of the waveguide as well as the flange attachment was inspected by Computerized Tomography (CT) scanning. This inspection step was performed at key points during the test flow and at the end of the test program. This allowed close examination of the exact cross -section of interest and was very e ffective in inspecting and ensuring the quality of the flange attachments. No defect of any kind was observed on waveguides with electroformed flange attachments at any point during the test progra m. CT-scan images of the flexible parts and a flange attachment are shown in Figure 11.

Document

AMS_2020.pdf

538 Figure 11. CT -Scan of Flexible Waveguide with Electroformed Flange RF Test: RF performance was evaluated by measuring the insertion loss and Voltage Standing Wave Ratio (VSWR ) across the entire WR -34 frequency range of 22 GHz to 33 GHz . This test was performed with the waveguide held in the straight and 45 degree bent positions . Typical RF performance results are shown in Figure 12. Figure 12. RF Characterization of Waveguide

Document

AMS_2020.pdf

539 Test Program Development Tests During the development phase, two waveguides were subjected to a very detailed deve lopment test program. The successful performance in these tests is what lead to the selection of waveguides for the PSP program. The development program was as follows: 1. Waveguide visually inspected at 80X ; no significant defects observed 2. RF performance ch aracterized in straight and 45 degree bent c onfiguration and was acceptable 3. Resistance torque was measured; max imum resistance torque at 45 degree was less than 30 mN -m. 4. Waveguides were subjected to 4 00 flex cycles at ambient temperature between -45 degree and +45 degree and following this cycling, were visua lly inspected and characterized a. No change in resistance torque after life cycle test b. Visual inspection at 80X showed no new damages or degradat ion of existing surface blemishes c. No changes noticed in RF performance for either article 5. Waveguides were subjected to sine survey and random vibration; f ollowing vibration testing, the waveguides were visually inspected and characterized a. Visual inspection un der 10X did not show any issues 6. Waveguides exposed to two thermal cycles between - 125°C and +100 °C; at each temperature extreme, waveg uides were subjected to 100 bend cycles between - 45 deg and +45 deg a. Visual inspection under 10X did not show any issues for either article b. No changes in RF performance for either article Batch Qualification The flight batch of the wavegui des were subjected to a test program consisting of batch qualification test on one waveguide and acceptance testing on all flight and spare components. The first round of tests was performed on waveguides with soldered flanges and resulted in failures of some of the soldered joi nts as described earlier . The program described in this section was performed on the waveguides with electroformed flanges. Prior to the delivery of the waveguides the manufacturer performed 7 thermal cycles between -105°C to +170°C and RF characterization. The qualification program was as follows: 1. Visual Inspection (20X to 80X) of exterior and inspection of silver coating on interior with borescope 2. RF performance test 3. Radiography 4. Vibration Testing - sine sweep, sine vibration and random vibration (Protoflight) a. Visual Inspection showed no damage b. No changes noticed in RF performance post vibration 5. Bend resistance measurement and 330 flex cycles over ±45° (110 @ ambient, 110 @+170°C and 110 @ -105°C) a. Visual Inspection sh owed no damage b. No changes noticed in RF performance post mechanical function test 6. Thermal Cycling (200 cycles between - 105°C and +170°C) a. Visual Inspection showed no damage b. No changes noticed in RF performance over the course of thermal cycling 7. Radiography - no defect observed post thermal cycling (Figure 10) 8. Removal of side wall and visual inspection of the interior; silver coating was found to be in good condition (Figure 10) Acceptance Test Five flight waveguides were subjected to the following tests: 1. Visual Inspection (20X to 80X) 2. RF performance test

Document

AMS_2020.pdf

540 3. Radiography 4. Vibration Testing – sine sweep and random vibration (Acceptance) 5. Bend resistance measurement and 5 flex cycles, ±45 degrees @ ambient t emperature 6. Visual Inspection (20X to 80X) 7. Radiography 8. RF performance All waveguides successfully passed the acceptance test program. Two of the waveguides were installed to the flight HGA assembly. Figure 13. Waveguide Placeholders Locate Bulkhead Flange Relative to Feed HGA Assembly Flight Installation The flex waveguides span the gap between the waveguide bulkhead on the spacecraft side and the antenna feed assembly on the HGA side, across the axis of the drive actuator. The antenna feed assembly is installed at the center of the main reflector, which is mounted to the actuator. The bulkhead feedthrough connects the HGA to the RF components located on the inside of the spacecraft . Though FWG s have less stringent alignment requirement than a RF rotary joint, they should be subjected to pure bending only and any extension or twisting must be avoided. Therefore, the feed ass embly and the bulkhead feedthrough must be aligned and the distance between them must be controlled precisely to match the length of the waveguides. T he two flange interfaces were both positioned at equal distance from the drive actuator axis. The alignment requirement was met by performing this installation on a tooling platform using waveguide Placeholders. The flexible waveguides are installed to the HGA assembly with the actuator at the 0-degree position where both waveguides are straight without bending. The HGA - actuator assembly and the waveguide bulkhead br acket were located on the tooling platform, and the two waveguide Placeholders were installed to the feed assembly on the reflector. Shims were added under the waveguide bulkhead as needed to allow proper mating to the Placeholders. The waveguide placeholders were then replaced by the FWG one at a time (Figure 13). The placeholders were machined to high precision to match the FWG length and bolt pattern. The pinned- halves design (Figure 14) allowed easy removal of the Placeholders to be replaced with the ac tual waveguides .

Document

AMS_2020.pdf

541 Figure 14. Exploded View of Waveguide Placeholders To maintain the relationship between the bulkhead waveguide and the reflector assembly, a handling fixture was used during transportation and handling (Figure 15) . All mounting interfaces between the HGA assembly and the tooling plate or handling fixture were designed with a nominal shim that could be adjusted as required during installation. The detailed integration and handling procedures using the associated tooling and fixtures ensured that the waveguides were installed with utmost care to avoid undesirable loading. Figure 15. Shimming Opportunities Lessons Learned Several valuable technical, procedural and philosophical lessons were learned over the course of this task . Following is a brief description of the key lesson learned: • Electroformed flexible waveguide made out of Ni -Co alloy, is a viable alternative to RF rotary joints and coaxial cables, under certain situations. • End flange attachment technique has a significant influence on RF losses. • All aspects of new product should be thoroughly examined, not just the characteristics of interest. While our program was primarily focused on the flexibility and life cycle, the only issues we faced were related to flange attachment. • Early testing during development phase greatly enhances confidence. • The t hermal cycling test is very critical to evaluate flexible waveguides. Detailed thermal analysis can be very effective in predicting the effect of RF transmission on waveguide temperature.

Document

AMS_2020.pdf

542 • Superior technical options when available should be given due consideration and not be dismissed just due to lack of heritage. • Laser vibrometers and miniature teardrop accelerometers can be effectively used in vibration testing of flexible parts without influencing dynamic performance. • Bore scopes can be used to inspect silver coatings on the inside of waveguides; this was valida ted by RF tests and destructive inspection. • CT scan is an effective tool to inspect soldered flange joints. RF performance can be directly correlated to defective joints observed by CT scanning. Figure 16. HGA Assembly on PSP Spacecraft Summary The PSP HGA assembly successfully used two flexible waveguides for RF transmission across the actuator rotary joint (Figure 16) . The analysis and test programs were effective in proving the suitability of the FWG for on- orbit cyclic o perations. They were also effective in detecting some lat ent process defects and resulted in the use of an improved waveguide with superior RF performance. The PSP spacecraft was launched in August 2018 and has completed three close encounters with the sun so far. All aspects of the HGA and the RF system are performing well as planned. Acknowledgement The authors would like to thank the members of the JHUAPL Space Simulation Lab for their support of this task over a period spanning several years , Perry Malouf of JHUAPL for performing RF test s, and finally Clency Lee Yow and his team at Custom Microw ave, Inc. for their support.

Document

AMS_2020.pdf

543 Thermal Vacuum Testing Lessons Learned for Small Stepper Motors and a CubeSat Translation Mechanism Alex Few*, Lynn Albritton* and Don McQueen* Abstract Near Earth Asteroid (NEA) Scout is a deep space satellite manifested on NASA’s ARTEMIS 1 as a secondary payload. The spacecraft meets the CubeSat 6U standard (about 300 x 200 x 100 mm) and is designed to travel 1 AU (150,000,000 km) over a 2- year mission t o observe a NEA 1. Once dispensed from SLS, the NEA Scout will use an 85- m2 solar sail to maneuver from lunar orbit to the asteroid. One of the critical mechanisms aboard NEA Scout, the Active Mass Translator (AMT), serves as a trimming and momentum managem ent mechanism for the sail system as it balances the sail center of pressure and the vehicle’s center of mass . The AMT produces 150 x 68 mm of translation at sub- millimeter precision and accommodates a shielded wire harness and coax cables during operation. The system has strict power, mass and data budgets and must survive operation in a shaded deep- space environment . The AMT system has recently completed and passed environmental testing. This paper will discuss lessons learned through three consecutive thermal vacuum tests spanning nine months and includes insight from the NASA Marshall Space Flight Center (MSFC) design/test team, NASA MSFC Subject Matter Experts in DC motors and electronics and the NASA Engineering and Safety Center (NESC) Mechanical Syst ems Discipline Team (MSDT). Important points of discussion will include (1) failure modes of a micro stepper gear motor in vacuum and destructive analysis findings, (2) instrumentation of a TVAC test to determine stepper motor health in near -real time, (3) determination of the duty cycle at a given operational environment, and (4) the design of the TVAC test profile to discover thermal capabilites of the micro stepper motors in vacuum. Papers were previously presented at the 44th Aerospace Mechanisms Sympos ium entitled “Testing and Maturing a Mass Translating Mechanism for a Deep Space CubeSat” and the 43rd Aerospace Mechanisms Symposium entitled, “Development of a High Performance, Low Profile Translation Table with Wire Feedthrough for a Deep Space CubeSat ”. Introduction The AMT was added to the Near Earth Asteroid Scout project late in the design cycle. So, the volume for the device had to be carved from other subsystems already under design and development. This late addition caused the design team to make design decisions that would not normal ly be recommended in order to have a chance at meeting volume, cost and schedule targets. One such decision was to use a commercially available space rated stepper/gear motor arrangement. This paper describes t he problems encountered in using this motor design in vacuum. and the steps taken to complete qualification testing of the AMT design. The AMT test phase was originall y scoped as a one week activity but grew to nine months due to back -toback stepper motor failures during TVAC testing. Prior to further discussion, It should be noted that the motor s chosen for the AMT design were the smallest motor s commercially available, and had been used in a flight design befor e. However, the previous flight proj ect app lication was in atmoshere (in cabin) promoting convective heat flow . Since this was a first -time in -vacuum application for these motors, significant design changes (discussed in detail in previous papers cited in the abstract) were made to create a more thermodynamically and mechanically robust design. The mechanical interfaces were revised to include an indium and aluminum clamshell design. This clamshell approach used all surface areas available on the motor and transmission casing as thermally conductive paths . Even so, both failures proved the * Marshall Space Flight Center, Huntsville, AL Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 2020

Document

AMS_2020.pdf

544 need for even more improvements, both to the hardware and test design. Nine months prior, the Engineering Development Unit (EDU ) system demonstrated successful, long- term operation between - 50 and 60°C. Needless to say, failures in s uccessive flight TVAC tests at 2 5 and - 35°C were terrible surprises. During the two failure investigations —the first lasting about 13 weeks and the second about 9 weeks —the design and test team met with s ubject matter experts from both MSFC and from the NESC MSDT to identify possible causes of the motor overtemperature failures, realistic remediations , and new methods to capture test data . The major events between December 2018 and September 2019 are illustrated in Figure 1. Each of these events will be discussed at greater length. Figure 1. NEA Scout AMT TVAC Test Phase Events AMT TVAC Test 1, Failure at 25 °C The flight NEA Scout AMT , shown in Figure 2 left, was assembled in December of 2018 to the exact specificatio n represented during the EDU TVAC tesing earlier that year . Notable differences include : 1. Stepper motor manufacturing lot. 2. Use of flight -like mechanical interfaces , shown in Figure 2 right 3. Use of flight spare Motor Controller Board (MCB) instead of a protob oard 4. Different power supply 5. Test chamber setup, including new LED lights, thermocouples and col d plate interface Figure 2. NEA Scout AMT and Test Configuration Following a functional test verifying the electrical, mechanical and test chamber set up, the TVAC chamber was controlled to 2 5°C, given time to dwell and depressurized to <1.0E -5 Torr . The AMT motors were powered to begin the first functional test at ambi ent temperature (25° C). About a minute later , the team noted temperature spikes from TCs mounted near the motor housings . After pausing the operation and reading the motor coil impedances, it was clear that the motors had begun the process of a cascading s hort failure. Given the previous TVAC tests for the EDU, the team had a natural inclination to suspect 2

Document

AMS_2020.pdf

545 overheating of the motor coils, but the root cause was unsure . Previously, it had been a poor thermal conductive path, which had been addressed for this mechanism. The next 13 weeks would revisit and uncover a few new contributers that were not discovered during the EDU tests. Investigation Method, Failure Mode and Destructive Investigation The team chose to maintain the test configuration for a time until a detailed plan was developed in order to preserve any contributions from the test facility and set up. The plan included (1) collection of all data, compiled and reviewed to hint at a “smoking gun”, (2) a controlled return to ambient, (3) detailed documentation of test configuration tear down, (4) return hardware to cleanroom for non- destructive evaluation, (5) removal of stepper motors from hardware, (6) precision x -ray of steppers, (7) destructive evaluation of stepper motor internals. Data logs were compiled and compared to previous temperature and chamber pressure data. A compiled temperature and pressure data set is shown in Figure 3. The data did show a temperature spike and a pressure rise in the test chamber correlating perfectly with a current spike from the power supply . The small pressure rise is indicative that the lacquers on the motor windings reached an overtemped state and began to offgas. Hundreds of pictures were taken throughout the process, the most valuable of which would come during the destructive evaluation of the failed stepper motors. Figure 3. NEA Scout AMT Ambien t Failure Temperature, Current and Pressure Data NASA MSFC has a team and facillity devoted to failure investigation and these resources were vital to the stepper motor destructive evaluation process . Figure 4 shows some of the sample images of the failed stepper motor . These motors, which are about 6 mm in diameter, showed clear signs of overheating at first glance. The discolored (reference the coil at 10 o’clock in the right image of Figure 4) and deformed (same coil shown in the right image at the top) exhibited tell -tale signs of the coils overheating. The overheating was a two- fold failure . First, the lacquer break down emitted material causing the noticable pressure spikes . Second, the rapid temperature rise causes the copper wire to lengthen, creati ng the wavy, deformed coil . Y Motor Failure X Motor Degrades Chamber Pressure and Current spike simulaneously as temperature rises accelerate.

Document

AMS_2020.pdf

546 Electrical testing would also show that the deformed coil was indeed the coil that failed, rendering the motor useless . The motor vendor, which was gracious to participate in these efforts, agreed to these conclusions. Figure 4. Motor Coil Images post Destructive investigation . Note waviness to coil in 3rd image, top coil Though the symptom of the failure was obviously coil overtemp, the investigation team was unsure why these motors would fail at 25° C in a minute when motors of identical design were used on the EDU and lasted for up to an hour at 60° C without issue. Unfortunately, due to scheduling constraints, the program required the rebuild the AMT before the fundamental cause could be determined . The team chose to protect the motors from future failure by creating a “duty cycle” for the motors at 25° C and above. This duty cycle would be determined by the rebuilt hardware in the TVAC facility at vacuum . A sequence of increasingly long run times followed by coil impedance measurements would determine when the coils would reach a maxiumum allowable temperature. The impedance would then be monitored until it returned to a near - ambi ent temperature, allowing for the motor to be cycled again. Each mo tor coil was monitored, and the performance ranges between coils and motors were larger than expected. The first lot of duty cycle discovery gave a maximum of 40 seconds on, 25 seconds off . This duty cycle would keep the motors from reaching a “red line” test limit of 110° C maximum coil temperature. The team chose to use the duty cycle operation at the 25° C and 45° C functional tests. Lessons Learned from AMT TVAC Test Failure 1 1. Pressure, tem perature and current data should be collected at a much faster rate than we previously thought during TVAC testing involving DC motors. These motors weigh less than 4 grams, and the coils may be a few tenths of a gram each. Their overheating still produced a measurable pressure spike of nearly 200% above the ambient pressure at 0.5E-5 Torr. One second data rate was chosen for further testing. 2. Thermocouples should be place d on available motor surfaces and a nearby heat sink to estimate thermal conductance. Close correlation of the two temperatures are a great sanity check that the system’s heat is conducting as expected. 3. This motor coils’ performance has a wide range of vari ability . You cannot base performance of a system on a previous motor set . (The team would later discover that within a single motor, one coil would require a hefty duty cycle of 50% while the other didn’t need one at all . The team b elieved this to be l argely because the motors were not or iginally designed for a vacuum application. ) AMT TVAC Test 2, Failure at - 35°C The Flight AMT w as rebuilt with new flight motors and returned to the TVAC facility in June of 2019. Using the newly employed duty cycles at ambient and hot temperature ranges, the team felt confident that the issue was resolved using the operational change. The ambient and hot functional tests went without issue . The hardware then was tested at the cold extreme of -35°C. After a 20 -minute continuous operation, the motor current spiked— indicative of a short in the windings due to overtemperature— and the test was aborted. The failure at the second protoflight TVAC test was completely unexpected and gave more insight

Document

AMS_2020.pdf

547 into the motors’ sensitivity to thermal environments and conductive paths . Prior to the second test, the team determined a duty cycle for the motors at the ambient and hot extreme. Failure Mode, Preliminary Findings and Destructive Investigation Since the first failure, the test plan was updated to use a data logging power supply to monitor current and voltage levels across the motor driver circuits . Analyzing these log files, the team found some peculiar current behavior which resembled a motor failure on the Tethered Satel lite mission. (The motor on the TSS mission was determined to have failed due to a phenonmenon called Paschen Discharge.) This current behavior is shown in Figure 5. Since the faliure occurred at such a low temperature, the team—believing the causes of fai lure may have been different than before—spent weeks investigating Paschen Discharge. This effort was never fully realized due to schedule constraints, but the team could never rule out the possibility completely . The concluded hypothesis was that the heat ed coils offgassed small amounts of lacquer which filled the small cavity inside the motor . The motor vent paths, sealed by kapton tape and/or indium, disallowed any free molecules to escape and possibly raised the pressure into the critical 10-4 to 10-3 Torr range. This pressure, coupled with the local magnetic field may have provided a “sweet spot” for the discharge to occur across particular lengths of winding. This hypothetical event could have caused the currrent to rise to 0.7 amp, hold for a moment, and once winding temperatures exceeded a failure temperature, create a short. Figure 5. Potential current rise and pause could be Paschen Discharge The failed motors from the second TVAC test w ere also taken to the MSFC failure analysis team. This second investigation yielded some helpful data in regards to the manufacuring variations in the motors, coils, magnet wire and potting . The motors underwent a typical round of X -ray imaging, which plainly showed deformed coils, as shown in Figure 6. The destructive evaluation showed a few new irregularities . One of which showed that the ball bearings inside the motor were single shielded and were installed with no regard to the shield direction. The next i rregularity was with the laquer . A special instrument allowing metals and polymers to be clearly distinguished under microscope showed that the laquer was producing

Document

AMS_2020.pdf

548 very large bubbles/voids in random areas . Figure 7 shows an image of a few bubbles or voids formed in the lacquer, either due to testing or remnant of manufacturing processes . This image, amongst others collected, also gave insight that the coil/stator design was not suitable for sustained use in vacuum. The most prominent detail noticed at this investigation w as with the potting and mounting of the coils to the stator . The team assumed the coils were potted to the stator, but a coil was accidentally damaged and became detached from the stator . Upon further analysis, it became clear that each coi l was potted to the stator in a single location, with a single dab of adhesive which was not especially thermally conductive . An example of this potting location is shown in Figure 8. Figure 6. Deformed Stepper Mot or Coil under X -ray Figure 7. Coils under high magnification showing some evidence of voids in insulation

Document

AMS_2020.pdf

549 Figure 8. Potting location for coil to stator The potting underdesign and presence of voids in wire insulation were clear signs that these motors were not well suited for a vacuum environment operation. The conductive paths from the windings to the motor casing were undersized, and the magne t wire had thin insulation. These findings paired with the manufacturing variance of the potting and insulation thickness made it clear that the motors would have to complete a more enhanced thermal acceptance testing to choose the more robust motors from a particul ar lot. Only after completi ng these tests could the motors be integrated into the flight hardware for an AMT subsystem TVAC test . The details of the acceptance and burn- in testing will be covered in the next section. Lessons Learned from AMT TVAC Test Failure 2 1. Motors intended for vacuum use should use coil wire with thicker vacuum compatible insualtion. Some vendors will twice or thrice dip their products for vacuum application. 2. Magnet wire defects such as thin insulation and voids/bubbles can be susceptible to first timewide temperature swings (such as a first time run in vacuum). 3. Motor coils should be potted on all available surfaces to maximize thermal conduction . These motors, though advertised as vacuum rated, were contacting th e stator in a single, potted location. This potting was analyzed and had little to no thermal conductive properties . The team’s thermal analyst would have assumed these coils to be relying on radiative cooling alone. 4. When using small (<15 mm) DC motors, or any motor from a non- flight rated vend or, a detailed motor thermal model should be developed to estimate a steady state temperature or a first cut duty cycle . Buy some motors, break them open, and figure out how the heat is conducting to the surface. 5. Acceptance testing in vacuum at the motor level is a must. Again, buy extra motors in case of failures. 6. Motors may be susceptible to discharge- related failures if the insulation offgasses due to rising temperatures. Motor Conditioning and Burn -in Test Met hod An initial effort was made to determine a method to recreate a Paschen Discharge. Exploring the variables that factor into achieving this effect led to the realization that the conditions could not be duplicated. Focus turned instead to preventive tac tics. Ultimately every motor failure traced back to coil temperature, either by causing off -gassing that created an atmosphere to carry a discharge, or by outright melting the wire coatings, resulting in a short. This points to the wire coatings as the weakest link, and coating inconsistency between batches as one of the hurdles. A two- part approach was planned. First, develop a conditioning and vetting test series . Second, determine duty cycles to prevent coil overheating.

Document

AMS_2020.pdf

550 The conditioning tests were designed to gradually “burn in” the motors by incrementally heating and allowing to cool in vacuum. Coil temperatures would gradually approach but be kept below manufacturer’s rated value. This would theoretically allow “bubbles” in the wire coatings to work their way out gradually through softened coating material without leaving permanent damage. A duty cycle is determined as the amount of time the motor is allowed to run in a specific environment before forcing a rest/recovery period. The run and rest times were, in this case, selected to limit the coil maximum temperature, and cool to a minimum temperature selected by the team before running again. Determining duty cycles were actually a by -product of the conditioning test series, and were established and r ecorded along the way. This testing would also demonstrate each motor’s individual capability to survive operation in a vacuum environment. Overall performance during the test series would be used as a basis for ranking flight motor candidates. A sample of the motors final duty cycle is shown in Figure 9. Figure 9. Duty cycles for AMT flight motor candidates at 25 °C and 35° C. The four motor candidates shown above were from the exact same vendor lot and were run with the same max allowable coil temps and cool down targets, but the duty cycles ranged from 36% to 57% at 25 °C and 30% to 48% at 35° C. The previous use of thermocouples to monitor coil temperature was insufficient for closely monitoring the coil temperature due to the thermal lag throu gh the motor casing, indium, and clamshell bracket to the thermocouple. Instead, the team turned to resistance measurements. Coil temperature correlates to motor impedance. Resistance of the coils was measured while the motor was off, and immediately being energized. Resistance- based temperature measurement was calculated for each coil, and verified against the chamber temperature prior to start of testing. More on this method is discussed in “Testing and Maturing a Mass Translating Mechanism for a Deep Space CubeSat” from the 44th Aerospace Mechanisms Symposium. For all motors not previously used in a vacuum, a standard vacuum bake- out was performed to reduce the amount of material that would offgas during the duty cycle and burn in testing. Test plan Three vacuum chamber temperatures were selected for duty cycle determination: room temperature (25°C), max operational tempature (45° C), and a mid value (35°C). Motors were mounted in a flight -like manner , wrapped in indium and installed in EDU chassis; no drive train hardware was included, allowing the motors to freely rotate. Maximum coil temperature and minimum cool -down temperatures were selected, and resistance was determined corresponding to each temperature. Two voltages were selected, the first being at the low -end of controller capability, the second at expected mission level. In this mission case, it was 9.8- 12.0 V. Multiple voltages were selected to test the sensitivity of the coil maximum

Document

AMS_2020.pdf

551 temperatures to voltage supply . During testing, the team n oticed that the motors actually ran hotter at a lower voltage. This was because the motor driver would send a higher current at the lower voltage, which was counter to intial assumptions. The assumed most favorable conditions were selected for the initial test, being lower voltage and coolest temperature. A duty cycle was determined, and then a total of 15 minutes of run time was performed at that duty cycle. The voltage was increased and the test repeated. Then chamber temperature was increased, and the t est repeated for both voltages, and the entire series run again at the third chamber temperature. A duty cycle was recorded for each specific motor at the given voltage and chamber temperature in which it was determined, so that each motor had six duty cyc les established. At the end of this series, the motor was considered “conditioned”. Duty Cycle D iscovery Method • Run motor for a few seconds • Measure and record a ll coil resistances immediately after operation ends . Recommend to have a multimeter in the loop of the driver circuit. • Continue to monitor resistance until reading is below the cool -down value. o Note that cooling in vacuum is asymptotic in nature. o Cooling to within 5- 10°C of ambient gives a good restart point for the following cycles • Run motor a few seconds longer than the previous run , depending on motor performance trends. o If the motor max coil temp is below maximum allowable, the following cycle can be longer o If the motor max temp is met or exceeded, either reduce the run t ime or extend the cool down time. Again, note that extending the cool down time is more time expensive than reducing the on time. • Continue until three runs are made with the same runtime, cool -down, and resistance values. This indicates a steady state duty cycle. A sample of the Duty cycle discovery data is shown in Figure 10. The data shown in Figure 10 is extensive, but still reduced from the total data set (a sample of which is shown in Figure 11) the team used to finalize a duty cycle for flight operations. The most pertinent data is the red and green boxes, which show the increasing run and cool down times as the approach a final value. For the left sample, this time was about 20 second run time, followed by a 55 second cool down. The right sample had a 35 second run time and a 70 second cool down. The teal line shows the duty cycle as a percentage of run vs. cool down time. Note the differences between the two identical motors, further indicating some thermal differences resulting from manufacturing v ariance. A final data point of note is the light green line, showing the current draw during a duty cycle. The upward sloping nature during each on cycle is indicative of the heating coil. As the coil temp rises, the resistance does as well. The motor driv er hardware sends more current as a response. The peaks of the green line during a cycle could be charted to give a maximum coil temp trend. The peaks remained horizontal for this data set, hinting that the coil temperature maximums were holding constant across the cycles. If the peaks were rising gradually, the coil temp could be assumed to be rising as well and the duty cycle could be adjusted accordingly. The same is true if the peaks were falling, except the coil temps would be decreasing and the duty c ycle could be adjusted in kind.

Document

AMS_2020.pdf

552 Figure 10. Duty Cycle Data for two Stepper Motors Motor Burn- in Following a duty cycle discovery run in a vacuum environment, the motors were run per the discovered duty cycle until 1 5 minutes of motor run time had occurred. Resistances and times for run and cool -down were recorded to verify that a long duration cycling would not overtemp the motor coils. This burn in time was used to help coil insulation material offgas at a slower, more controlled rate, and reduce any opportunity for an arcing event to occur . Further, the burn in time would reduce the thermal shock to the insualtion and reduce the risk of voids forming in the insulation. AMT TVAC Test 3, Passed The final flight TVAC test plan included even more data capturing requirements: (1) Local temperatures on AMT logged at 1 Hz, (2) Chamber control temperature taken logged at 1 Hz, (3) Chamber pressure logged at 1 Hz, (4) Motor circuit current and voltage data logged at 5 Hz, (5) Control Board Memory and Co mmands logged at 1 Hz and (6) Motor Winding impedances taken immediately after operation termination. These data sets were combined into a master data set where temperature, current, pressure, and operation speeds could be compared at each thermal cycle. An example of the combined sets is shown in Figure 11.

Document

AMS_2020.pdf

553 Figure 11. Combined data set . Note orange line documents X axis change in position. Duty cycle = 40%. The extensive measures taken to understand the details of the motors’ internal design paired wit h the meticulous process of duty cycle determination and burn in testing yielded a long- awaited succesful test . The third iteration of the NEA Scout AMT test passed with flying colors . The motors, though unable to perform in vacuum under typical operation, showed an ability to operate in vacuum for thousands of duty cycles and an accululative run time of over 10 hours . All data points were reviewed in detail following the conclusion of test 3. A single engineer took a few weeks to align time signatures ac ross each data source and compile the data to create tables and charts similar to Figure 10 and 11. The data was reviewed and compared with the thermal environments in the test chamber to inform the final control parameters for the flight motor controller boards . These data sets will also be valuable to compare with flight data in the event AMT operation telemetry is reviewed or modified . Conclusion In conclusion, the AMT system TVAC test failures caused the team to develop new processes to determine a motor’s duty cycle prior to environmental test with limited risk to flight hardware, instrument and collect data for post -test analysis, and perf orm destructive analysis on failed micro gear motors. The motors used on the NEA Scout AMT were determined not to be suitable for 100% duty cycle in vacuum, but a conservative duty cycle enabled the AMT to operate and meet long- term mission requirements . The AMT motors were instrumented to determine motor health in near real time. Previous failures during NEA Scout AMT TVAC test always pointed back to an unfit thermal conduc tive path, most likely due to under -insulated magnet wire and poor thermal conduction from motor coils to a larger heat sink . It was hypothesized that Paschen discharges could occur in a small stepper motor such as these if the coil temperatures rises and offgasses enough material, though this hypothesis could not be recreated or confirmed through further testing.

Document

AMS_2020.pdf

554 Major Lessons Learned 1. A Commercial off the Shelf (COTS) non space- qualified motor could be used for flight purposes if proper motor conditioning and duty cycling is determined in a comparable space environment 2. Any COTS non space- qualified motor should be destructively analyzed prior to use to accurately create a thermal model. 3. TVAC tests including DC motors should log motor temperature, power supply current, power supply voltage and chamber pressure at 1 second resolution, at minimum. 4. Motor internal pressure could increase to a pressure range suitable for arcing, discharging or ionizing if magnet wire insualtion breaks down due to overtemperature conditions, leading to failure. References 1. McNutt, L.; Johnson, L.; Clardy, D.; Castillo -Rogez, J.; Frick, A.; and L. Jones. “Near -Earth Asteroid Scout.” AIAA Space 2014 Conference; 4- 7 Aug. 2014; San Diego, CA; United States. 2. Few, A.; “Development of a High Performance, Low Profile Translation Table with Wire Feedthrough for a Deep Space CubeSat .” 43rd Aerospace Mechanisms Symposia; 4- 6 May 2016; San Jose, CA.; United States. 3. Few, A.; Lockett, T.; Wilson, R.; Boling, D.; and Loper, E.; “ Testing and Maturing a Mass Translating Mechanism for a Deep Space CubeSat.” 44rd Aerospace Mechanisms Symposia; 1618 May 2018; Cleveland, OH.; United States.

Document

AMS_2020.pdf

555 Design and Development of the GPM Solar Array Drive Assembly, Orbital P erformance , and Lessons Learned Alejandro Rivera*, Glenn Bock*, Alphonso Stewart**, Jon Lawrence**, Daniel Powers**, Gary Brown** and Rodger Farley+ Abstract The Global Precipitation Measurement (GPM) core spacecraft is a NASA/JAXA joint mission launched in 2014 . Developed in- house at NASA GSFC, it s deployable appendages include two large solar arrays each driven by a single axis solar array drive assembly and a gimbal equipped high gain antenna. Lessons learned from the Tropical Rainfall Measuring Mission (TRMM) – Y Solar Array Drive Assembly ( SADA) anomaly and Lunar Reconnaissance Orbiter’s (LRO’s) thermal testing, influenced the design of the GPM Solar Array Drive Assemblies . This paper describes the TRMM anomaly, design and development aspects of the GPM SADA, its on-orbit health and performance, and finally techniques and orbital maintenance maneuvers followed by the Flight Operations Team to minimize drag, actuator cycles , and extend the life of the actuators and the mission itself. Introduction The GPM core spacecraft is a joint mission between the National Aeronautics and Space Administration (NASA) of the United States and the Japan Aerospace Exploration Agency (JAXA). It is a Low Earth Orbit (LEO ) spacecraft launched on Febru ary 27, 2014 with a circular 407 km altitude and 65- degree inclination orbit selected to provide full global precipitation coverage updated ev ery 24 hours . GPM succeeds the TRMM spacecraft launched in November 1997 that measured precipitation over tropical and subtropical regions, from the Mediterranean Sea (35° north latitude) to the southern tip of South Africa (35° south latitude) . Measurements from GPM core, however, provide even greater coverage—between the Arctic Circle (65° north latitude) and the Antarctic Circle (65° south latitude). These measurements, combined with those from other polar -orbiting satel lites in the GPM constellation, currently provide global precipitation datasets every three hours. This integrated approach and unified dataset helps advance scientists ' understanding of Earth's water and energy cycle, produces enhanced forecasts of hurricanes, floods, and droughts , and helps enhance Earth’s climate models . Figure 1. The GPM Core (left) and TRMM (right) Satellites ______________________ ___ * KBR Inc – Space Engineering Division, Greenbelt, MD ** NASA Goddard Space Flight Center, Greenbelt, MD + NASA Goddard Space Flight Center, Greenbelt, MD (Ret.) Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center , 2020

Document

AMS_2020.pdf

556 The TRMM SADA Anomaly The TRMM satellite mission ended on April 8, 2015 after more than 17 years of very successful operation. While the mission clearly exceeded its original three- year mission goal, it experienced a minor anomaly involving one of the So lar Array Drive Actuators . TRMM had two deployable solar arrays designe d to track the sun with help of a single axis SADA . A Schaeffer Magnetics rotary actuator identical to that used on NASA GSFC’s X-Ray Timing Explorer (XTE) mission drove the TRMM SADA, and High Gain Antenna (HGA ) gimbal . The r otary actuator was a Type 5 ½ (T ype 5 modified with an output bearing from a T ype 6 drive) . It had a 3- phase, 6- state stepper motor . The harmonic drive was a “silk hat” type with a pitch diameter of 2.5 in (5 cm) and a 200:1 reduction. The flexspline was made out of 304L S tainless Steel (SS) , the circular spline and wave generator used 17- 4 PH SS and 440C SS respectively . Penn zane 2000 synthetic hydrocarbon oil with 5% lead naphthenate additive was used as lubricant on the gear teeth and bearings . An internal r otary incremental encoder with three absolute positions provided position, velocity, direction of travel, and was capable of being used in open or closed loop mode [1 ]. Figure 2. The TRMM SADA The SADA also included a cable wrap used to transfer signal and power across the rotary joint. Seventy - six 20 gage wires were sewn together to form two belts that spiral around a central reel . The belts were 36 in (91 cm) long by 2.5 in (5 cm) wide and were separated using 0.005- in (27 µm) Kapton to reduce sliding frict ion. The cable wrap had a maximum travel of ±300 degrees but the designed use was ±175 degrees where it could operate in the region of low friction. Finally, the TRMM SADA included a main deployment / shoulder hinge bolted onto the output face of the rotary actuator, which rotated the w ing 90 degrees from the spacecraft body with the help of constant torque springs and a rotary viscous damper kept warm with st rip heaters and thermostats [1].

Document

AMS_2020.pdf

557 Figure 3. Sun’s View of TRMM SADA at high beta angle An initial TRMM/XTE SADA rotary actuator qualification unit partial 1000- hr life test revealed that no signs of deterioration could be detected by any of the tests designed to monitor health (threshold voltage, output torque) . But post -test inspection revealed significant amount of wear in the bearing to flexspline interface and the gear teeth. The 304L flexspline inner diameter had galled with the 440C bearing outer race, the inner race had slipped down the wave generator plug, and the Pennzane 2000 oil was no longer present [1]. As detailed in Reference 1, a series of changes were made and then a combined TRMM/XTE SADA life test that covered 1.1x full TRMM life (11 million output degrees of travel) was successfully completed. The TRMM –Y side array, was always on the sun, or warm side (TRMM performed routine yaw maneuvers to keep one s pacecraft side toward the sun) and experienced temperatures greater than li fe test for substantial periods, while the +Y SADA on the cold side of the spacecraft, operated within predicted limits. Unfortunately, there was no operational way to reduce temperatures as the major heat input was from direct sun on the –Y SADA output flange and via conduction from the inboard boom and shoulder hinge. The excessive temperature resulted in increased (exponential) vaporization of l ubricants, which are required for good lubrication of bearings and harmonic drive for proper actuator life. The original 1000 -hour life test had proven that internal degradation may not b e apparent from telemetry, i.e., there may be no warning until catastrophi c failure so SADA could halt at some random position without notice. In September 2002 the –Y SADA showed first indication of impending failure as it was stuck during slew -to-feather for ~ 2 minutes . Prior to jamming, the SADA was also missing steps every 14 seconds during normal tracking (Figure 5) . The TRMM team then decided to discontinue sun tracking and park the – Y SADA at the feathered horizontal position as a precaution to avoid the possibility of that array becoming stuck in a non- prefer red position, which could hav e resulted in excessive drag . Figure 4. –Y SADA operated hotter than expected during mission

Document

AMS_2020.pdf

558 Figure 5. –Y SADA Anomaly: Prior to jamming for ~2 min, actuator was missing steps during tracking This lack of sun tracking with the - Y solar array led to slightly less available pow er, but still allowed sufficient power for nominal operations of all working instruments . In 2003 , the TRMM team had to decide whether the +Y SADA could reliably be expected to continue operating normally until fuel considerations forced controlled reentry to be initiated. An anomaly investigation resulted in thorough lubricant loss, thermal and life analyses being performed. Key aspects taken into account included the fact that the most cri tical temperatures are at the harmonic drive wave generator bearing / flex spline / circular spline interface, and the outboard motor bearing. Thermal analysis assumed thermal conductivity value across ball bearings taking advantage of oil meniscus spanning each ball/race interface. As lubricant is drive n off, meniscus supports less heat flow, creating higher thermal resistance and thus higher temperatures. Figure 6. TRMM SADAs Lubricant Loss Analysis This can lead to a runaway condition where balls eventually run dry. It is then possible that the –Y SADA lost lubricant and the dry running may have generated particles . Stainless steel wear particles generated can become ferro- magnetic due to work hardening. They can t hen be attracted to motor magnets and may clog the small air gap between coils and magnets . Without torque multiplication at the output, the motor may permanently jam. Another source of possible failure was the precipitation of lead naphthenate high-

Document

AMS_2020.pdf

559 pressure additive, which was detected to some extent in the life test . This effect may be more related to solubility and wear cycles. The actuator missed steps every ~14 seconds corresponding to every 180° of motor rotation which may indicate lead precipit ate at the interface of wave- generator bearing and flex spline that engages twice per motor revolution as each lobe passes by . Figure 7. +Y SADA Life Prediction Life analysis predicted that the +Y TRMM SADA c ould reliably be expected to continue operating normally until fuel considerations forced controlled reentry to be initiated. Analysis predicted that the 1.1x full TRMM life (11 million output degrees of travel) would be reached in March of 2006 and that 16 million degrees would be ac complished by July of 2010. The +Y drive operated well within temperature limits and did not experience the same problems. However, because of the situation with the –Y side solar array, the power subsy stem required special attention during State Of C harge (SOC) periods, especially during periods of low Beta angle. Eventually the +Y SADA operated successfully for a total of approximately 21 mi llion output degrees of travel , until the spacecraft was decommissioned on April 8, 2015. The GPM Solar Array Drive Assembly The GPM Solar Array Deployment and Drive System consists of two independent wings, each of which comprise five major elements. Shown in Figure 8, these elements include: a) Solar panel assembly consisting of a rigid panel that supports the solar cells with electrical wiring (cell, diode board, associated circuitry, temperature sensor) and coarse sun sensor on outboard panels ( quantity 4); b) Deployment boom assembly connecting the panel assemblies to the SADA output flange. There is one boom assembly per wing. The hinges within the boom assembly provide the deployment force to position the panel in the required orientation. The booms also support the harnesses trav eling between the panel assembly and the SADA; c) Restraint / Release (R/R) mechanisms securing the panel assembly and boom to the spacecraft in the stowed configuration. There are five Restraint/Release mechanisms per wing: four for the panel assemblies and one for the deployable boom assembly; d) Panel Hinges providing the deployment force for eac h panel assembly. There are t wo panel hinges per hinge line; e) S olar Array Drive Assembly (SADA) which rotates the boom and panel assemblies in order to track the sun through each orbit and provide means to transfer the power generated by the solar arrays and the sensor telemetry signals to the observatory.

Document

AMS_2020.pdf

560 Figure 8. The GPM Solar Array Deployment and Drive System The design of the GPM SADA was driven mostl y by structural, thermal, and life / Number of cycles considerations : structural as it had to help meet the 0.6 Hz minimum fundamental frequency for the largest Solar A rrays every built at NASA Goddard with a inertial load of 255 kg- m2; thermal as it had to ensure the actuator was kept within its acceptable temperature limits to prevent the issues that caused the TRMM – Y SADA anomaly we previously described; and life as it had a requirement of 17,500 cycles ~ 8,750,000°. The GPM Solar Array Drive Assembly consists of 4 main subsystems: electro- mechanical (rotary actuator), structural (actuator output & Harmonic Drive ( HD), spacecraft and hinge interface plates), thermal (radiators , hinge interface plate, and thermal tube), and harness management (cable wrap) . The rotary actuator designed and manufactured by Sierra Nevada Corporation’s (SNC) Durham, NC facility, is compris ed of a three- phase redundant stepper motor with a 200:1 reduction , a Size 25 T -Cup harmonic drive gearbox , with position feedback provided by integral redundant coarse and fine optical encoders. The actuators provide positioning within the SADA. The GPM SADA actuator has two hard stops limiting travel to ±125° and it uses a SuperDuplex™ thin section output bearing lubricated with Pennzane 2001- 3PbNp oil and 2000- 3PbNp grease. GPM’s structural, thermal, and life requirements were considerably more severe than previous NASA missions that had used this same type of SNC rotary actuator (SDO, LRO) . Figure 9. The GPM Solar Array Drive Assembly

Document

AMS_2020.pdf

561 On the structural side, the solar arrays had a 0.6 Hz minimum fundamental frequency requirement . Since the first mode of the array is cross- axis bending , it was designed to maximize the stiffness in that axis as much as practical . To this purpose, numerous technical interchanges with the actuator manufacturer were conducted to modify the actuator design in order to develop an actuator that could meet our requ irement for a bending stiffness of at least 339,000 N-m/rad. To maximize the bending stiffness, the output bearing was changed from a tradi tional duplex back -to-back pair with a set of spacers between the bearings to a Super Duplex™ bearing with a one- piece outer ring to minimize the number of components and sliding interfaces. The Super Duplex™ thin section bearing has a common ring, either a single outer (as in GPM’s HGA S and SADA actuator output bearings), or a single inner, which has two bearing paths in it . Because this type of bearing had never been used before on a GSFC mission, an evaluation and comparison of Super Duplex™ vs the standard duplex DB pair was conducted. Stiffness: Havin g a “one piece” rigid outer ring provides improved ring stif fness and reduced distortion, as there no longer is a two- piece assembly; Geometric Dimensioning and Tolerancing ( GD&T ): two key GD&T parameters that will affect the performanc e of thin section ball bearings are the radial and axial runout. The radial runout is a measurement of the thickness variation of the bearing rings. The outer ring is measured from the ball path to the outer diameter of the ring, and the inner ring is measured from the ball path to the bore. The radial runout is defined as the wall thickness variation of the rotating ring. The axial runout on the other hand, is measured from the ball path to the face of the bearing rings. The variation in thickness measured is the axial runout. The Super Duplex™ bearing has a common ring , which has two bearing paths in it. As these are ground at the same time, they “run- out” together. This helps smooth out potential torque variations that would be caused by the variations in axial and radial runout adding torque into the bearing; Assembly & Installation: the disadvantages of using two separable bearings are potential errors in misalignment and orientation during assembly. If however , a SuperD uplex ™ bearing is used, since the assembly is not separable, the possibility for installation orientation errors is eliminated [2]. Figure 10. SuperDuplex™ (left) vs. Duplex DB (right) [2] To maximize the bending stiffness of the actuator the internal contact angle of the bearing was increased by 50% and the internal preload was increased by 300% . The machined components that support the inner and outer ou tput bearing raceways were chang ed from Titanium to Stainless Steel for a hig her modulus of elasticity and the inner raceway support was thickened to add rigidity . Finally, to minimize the influence of bearing mounting clearances, the tolerances on the mating journal diameters were tightened, the minimum installation clearance was reduced, and bearings were matched with the mating machined components to minimize the clearance fit across all deliverable units and provide more consistent results . Detailed bearing analysis was performed to ensure the final bearing parameters and fit -up resulted in acceptable stresses for operating and static loading conditions . A detailed Finite Element Model of the actuator was correlated with preliminary subassembly stiffness testing to gain confidence that the final design configuration would meet ov erall actuator stiffness goals . The as -delivered actuators demonstrated bendin g stiffness in the range of 341,000 to 370,000 N-m/rad, above the target goal to meet the desired fundamental frequency . The increased contact angle and preload result in increased bearing friction and contact stress relative to the SDO/LRO heritage. Therefore, a successful 6.5x life testing was performed to verify the expectation that lubricant life would still be more than adequate for the mission. One of the main lessons learned from the TRMM anomaly to ensure that mission life requirements are met is that the harmonic drive materials

Document

AMS_2020.pdf

562 and lubricants must be properly selected . This was particularly import ant given that the life requirement of GPM SADA and HGA gimbal actuators significantly exceeded that of previous GSFC missions that had used the same actuator: SDO had a life requirement of 2500 cycles/900,000°; LRO’s was 5700 cycles/2million degrees; GPM’ s SADA is 17500 cycles/8.75million degrees and GPM HGAS is 48,600 cycles/18.5million degrees . To ensure the SADA and HGAS actuators would meet their life requirements Sierra Nev ada Corporation (SNC) Durham, NC facility and NASA GSFC performed a series of harmonic drive accelerated life tests with different combinations of circular splines and flexible splines materials in both ambient (2x life) and vacuum environment (100,000 cycles) [3 ]. An exceptional description of the testing process and results obtained is provided in Reference 3, and main results are summarized here. Based on discussions between Harmonic Drive LLC and SNC, the material combinations shown in Table 1 were selected for comparative life testing. All HD gear configurations were customized HDT -25, 200:1, T -cup component sets manufactured by Harmonic Drive LLC and all were of the same dimensional design. All combinations used the same wave- generator bearing material and design, and each was identically lubricated with Pe nnzane 2001- 3PbNp oil and Rheolube 2004 grease per standard procedures . For the ambient test (conducted in air at room temperature), the input to the harmonic drive gear was set at 500 revolutions per minute ( rpm) and a co nstant 2.93 N*m of friction load was applied to the output . The main results for this test were as follows: the two Melonite Harmonic Drives showed a reduction in torsional stiffness at the end of life and the presence of relatively large particles of metallic debris in the gear mesh and were eliminated from consideration. The two non -Melonite gears were relatively comparable . The 15- 5 vs. 15 -5 gear visually looked slightly better with regard to the circular spline teeth and gear lubricant but the Nitronic vs. 15- 5 gear was also acceptable. This latter one was selected as the least risky p ath forward due to concerns over how the 15- 5 vs. 15 -5 would perform during vibration and in a vacuum environment [3] . Table 1. Harmonic DriveTM Materials Selection Ambient (2x life) & Vacuum (100K cycles) Test Results [3] Based on the above results it was decided to perform a second round of testing for the two leading candidates but in a vacuum environment under thermal cycling. The vacuum test consisted of 100,000 cycles of reversing 180° output revolutions in a <5.0x10-5 Torr vacuum enviornment cycling between 0 and 40°C . Input to the harmonic drive g ear was set at 500 rpm and a cons tant 2.93 N*m of friction load was applied to the output . Results showed that innthis case, the vacuum environment had little to no effect on the wear rate of the two harmonic drives tested with small differences in wear between the units . The vacuum test confirmed the program’s decision to use the 15- 5PH flexible spline against the Nitronic 60 circular spline as the f inal configuration to be used on the GPM SADA and HGAS actuators [3]. As it can be seen on the cross sectional view Figure 9, the Spacecraft Interface Plate (SIP) is used to attach the Solar Array Drive Assembly (SADA) to the Lower Bus Structure panel . The design and analysis of the SIP was driven by the deployed fundamental frequency requirement of the Solar Array which translated in the need to have a very stiff SIP . The Cable Wrap attaches to the bottom of the SIP and both are stationary . The mounting flange of the actuator is attached to the bottom of the SIP and is also stationary . The Hinge

Document

AMS_2020.pdf

563 Interface Plate (HIP) on the other hand is mounted on the top of the actuator output and it rotates with it . The HIP provides an interface between the actuator and the shoulder hinge. I t blocks the sun from the actuator output, and it also provides some bending stiffness . The design of the thermal radiator from the structural point of view was driven by the minimum fundamental frequency requirement of 100 Hz. SADA structural design was both simple and effective. The solar array loads travel through the shoulder hinge and the HIP to the out put of the actuator , and then out to the lower bus structure through the SIP. The cable wrap inertial loads are transferred to the SIP via the cable wrap housing. The following load cases were analyized using finite element analysis: 22 Gs in X,Y,Z (Mass Acceleration Curve loads); 2050 N shear, 440 N axial, 290 N -m moment (Combined Qual Limit Loads); 55°C and - 15°C (thermal stress) . Stress analysis showed positive margins of safety on yield and ultimate for all components. Figure 11. Cross Section View of SADA Finite Element Model A superb thermal design was of critical importance for the succesful operation of the SADA which had to handle the thermal load due solar cell power, motor, and solar impingement . On top of the previously described issues experienced during the TRMM mission, the GPM SADA thermal design was also driven by lessons learned from the LRO mission . During the LRO observatory level testing it was discovered that the rotary actuators were ex ceeding their temperature limits as the power loses ( P=I2R) within the solar array harness were very significant and dumping heat direclty into the actuator. Modifications to the LRO SADA during and after thermal vacuum testing, such as adding radiators an d removing M ulti-Layer Insulation blankets still did not resolve the issues . This resulted in LRO having to make major changes to the on- orbit solar array operations to ensure functionality of the actuators . Furthermore, GPM had close to two times the solar array harness power as LRO and 3 times the mission life. Due to this , the GPM SADA underwent a significant design effort to accommodate the motor power, harness heat ing, and extreme environment . Table 2. Key thermal requirements The actuator design could not be modified due to contract with the vendor . Thermal design was then focused around the rest of the SADA components . Coatings and tapes were applied on the Hinge Interface Plate, Cable Wrap Housing / end plate, spools, as shown in Figure 12. A large radiator was attached to the

Document

AMS_2020.pdf

564 edge of the Spacecraft Interface Plate as shown in Figure 9. The cable wrap was modified to absorb / reject the heat generated within the harness: an aluminum thermal tube extends into the actuator quil shaft to isolate the harness from the quill shaft; the Cable Wrap housing is directly attached to the Interface Plate instead of the actuator’s accessory flange for direct heat path to the radiator; a 3- piece Titanium spool is used to isolate the Cable Wrap bearings from harness heat, minimize bearing gradients, and minimize thermal expansion loads into the Cable Wrap bearings; heat straps going from the Cable Wrap housing to the lower bus structure were added to help stabilize any transient temperature swings, and provide a stable, well-characterized heat sink; the Cable Wrap be arings use Pennzane lubrica nt instead of Braycote. This was consistent with the actuator bearings to ensure no cross contamination of lubricants . Also, Pennzane has more restrictive temperature limits but longer life than Braycote. Figure 12. GPM SADA Thermal Design The Cable W rap (CW) acts as the harness management device. It allows electrical & power signal wires to pass across the rotational joint with no relative harness motion external to the housing. The CW Mechanism is driven by actuator motion / rotation of the Solar Array . Both are connected using an Oldham type coupling with the female portion attached to the actuator quill shaft, and the male side attached to the cable wrap spool subassembly . This coupling design allows for positional, l inear, and angular misalingments between the CW and the Actuator and transmits torque without placing side loads to either shaft . Invidivual wires are sewn together to form two belts that spiral around a central rotating spool . This ribbon portion of the harness is clamped to both the rotating spool and the CW housing as shown in Figure 13. By “mechanically grounding” the ri bbon harnessing to the housing and rotating spool, in addition to “mechanically grounding” the round bundle to the S/A boom, all motion of the harness occurs within the Cable Wrap housing as “winding & unwinding” in a spiral motion about the rotating spool [4]. Figure 13. GPM Cable Wrap Transversal Section [4]

Document

AMS_2020.pdf

565 These two arm spiral wrap transitions to a round bundle to pass through the actuator quill shaft, and communicates 16 6 wires (132 power & 34 signal) across the rotary joint to the Solar Array . The two main belts are separated by a .25-mm ( 0.010 -in) thick sheet of Kapton w hich minimizes the sliding friction between the belts and their Dacron stiches . The GPM Cable Wrap has a rotation capability of ±178° (where it either winds completely on the inner diameter or out on the outer diameter) in excess of the ±120° requirement to minimize the bending stresses on the belts and allow it to operate in the ‘sweet spot’ region of low friction . Harnessing is clamped to a 3- piece spool assembly (for ease of manufacturing) made out of Titanium . The spool is supported by two radial bearings slip fit to the housing and spool end fittings . The forward bearing is axially constrained whereas the aft one is not, to allow for spool assembly tolerance stack- up and thermal growth. Bearing inner races are clamped to the spool end fittings and forward bearing outer race is clamped to the CW housing. Because there is a 20°C gradient across the bearing races a problem ar ises due to the differences in Coefficients of Thermal E xpansion (CTE) between the bearing races (Stainless Steel 440C), CW housing (Al -6061- T6), and the CW spool (Ti -6Al-4V) which required a 0.0005 – 0.0008 in (13 – 20 µm) radial play [4]. Figure 14. Cable Wrap Radial Bearings [4] Over large temperature gradients such as the ones that the CW experiences, the clearances between the components in contact with the bearing can be eliminated and excessive loads could be transmitted into the bear ing, loads which could potentially cause brinelling of the its races . For of this reason it is very important to make sure the right tolerancing is used at the inner and outer race interfaces . A thorough bearing thermal stress analysis was performed that took into consideration the compression of the bearings’ outer race due to the CTE mismatch (6061- T6 vs. SS440C) at this interface; contraction of the bearing’s outer race bore (where it contacts the ball ) due to this mismatch and associated loss of radial play; and loss of radial p lay across entire temperature range (0 – 60°C) due to a 20 °C temperature gradient between inner and outer race. Since there was interest in looking at possible interferences due to thermal growth because of CTE mismatches that could generate stresses on the bearing, the worst scenario for this condition was chosen: max. spool O uter Diameter (OD) vs. min bearing Inner diameter (ID) and min CW housing ID vs. maximum bearing OD . Analysis showed that even after loss of radial play the highest mean contact Hertzian Stress that the bearing balls saw was 6.5*108 Pa which was well below the 2.4*109 Pa that would cause brinelling on the races . SADA On -Orbit Performance The GPM Core Observatory as seen in Figure 8 operates in a near circular orbit of approximately 407 km with a 65- degree inclination . It is a three- axis stabilized spacecraft , nadir pointing for instrument observation of the Earth and its atmosphere, with the X -axis aligned with the velocity vector. Depending on Solar Beta angle, GPM flips 180 degrees in yaw such that it flies with either the +X or – X axis forward. However, since it has thrusters on both sides, it can execute maneuvers in both orientations without slewing. GPM’s orbit

Document

AMS_2020.pdf

566 was chosen to optimize science data capture for the platform’s Dual -frequency Precipitation Radar (DPR) and GPM Microwave Imager instruments and provide coordination with GPM Mission partner satellites . Both the Semi -Major Axis and Eccentric ity requirement tolerance allowes for the geodetic height (HGT) of the Core Observatory to be maintained within 397 km to 419 km for effective DPR operation and minimize the altitude variation per latitude crossing over the life of the mission. T he inclined (65 degrees) non- sunsynchronous orbit allows the observatory to sample precipitation across all hours of the day from the Tropics to the Arctic and Antarctic Circles, and expand on the observations performed by TRMM, the Core Observatory’s predecessor. The Flig ht Operations Team (FOT) manages the fuel usage based on an orbital maintenance plan derived from these parameters [5] . Table 3. GPM Core Spacecraft Orbit Parameters [5] Maintaining the orb it to these constraints required the FOT to perform orbit maintenance maneuvers as the orbit experienced changes due to predictable Earth perturbations and the impact of flying in a variable drag environment while optimizing fuel usage and actuator cycles . The first type of maneuver was a prograde delta- V maneuver in the velocity direction to boost GPM up into a higher orbit. This is referred to as a dragmakeup maneuver and is used to increase the SC’s altitude after atmospheric drag has caused orbit to decay down to the bottom of the required control box. The second is due to the eccentricity requirement. GPM has a tight eccentricity requirement such that orbit remains very near circular. They only way eccentricity can be fixed is by preforming a maneuver, however if the SC is already at the top of the required control b ox then it can’t go up any higher, and a retrograde maneuver must be done. These are done on an as needed basis . Solar activity is a key factor . During periods of high activity the atmosphere expands , resulting in higher drag and faster orbital decay. The opposite takes place when it is low . Figure 15: GPM Mission Solar Array Tracking Profiles vs Beta Angle Hence, per pre- launch analysis, the mission was expecting to plan and perform orbit maintenance maneuvers as frequently as once a week during high solar activity, and as infrequently as every eight weeks during low solar activity. I ndeed, for the first few mont hs post-launch, maneuvers were performed once a week in the higher drag conditions as predicted. The solar array tracking profile used during this period was Profile G as shown in Figure 15. During the on- orbit check -out phase the senior engineers determined that the solar arrays performance was more than adequate to recharge the batteries and provide power to the bus, thus a few of the solar array feathering profiles were tested on board to optimize the input power and drag induced. After a few months of testing, the low drag profile know n as Profile K was selected in August of 2014. This change, coupled with the solar flux dropping slowly as the mission months passed, allowed for the maneuver frequency to slowly change from once a week to roughly o nce a month [5] . The periods of extreme low solar activity with low drag allows for more orbital eccentricity growth between maneuvers. And for any one maneuver during this time (constrained by an upper altitude limit), the eccentricity cannot be reduced sufficiently . Hence the Drag Make- Up maneuvers required to maintain the Semi -Major Axis within tolerances were both too infrequent and of insufficient size to control the eccentricity .

Document

AMS_2020.pdf

567 The FOT needed to develop an option that would allow GPM to control eccentricity growth without violating the geodetic height requirements, or at least one that would minimize loss of science for the remainder of the prime mission. Solution proposed was to increase GPM’s drag area profile by unfeathering the solar arrays panels. Similar to how the array profile was changed during early orbit check out to reduce drag, the FOT could change the array profile to one that could create sufficient drag needed to l ower the spacecraft’s geodetic height faster than the growth of the eccentricity. The benefit would be that no additional maneuvers would be required to maintain orbit requirements. The predicted analysis showed that using the largest drag profile, profile G, the eccentricity requirement could be maintained for several months . Hence, as shown in Figure 16, in August 2016 the Tracking Pr ofile was changed back to G [5]. Figure 16. -Y and +Y SADA Tracking Profiles used since launch date When using Profile G, trending was showing the SA profile induced additional cycles on the SA actuators , thus putting an unnecessary strain and risk to the spacecraft. Pre- launch analysis predicted that 3- years into the mission the Solar Arrays would cycle (out and back) around 18,000 times per wing . This was seen as unnecessary and so a switch was made to profile K . Due to the switc hing from the Profile G to the Profile K control table early after launch the +Y gimbal had only reached just under 5000 cycles, and the -Y gimbal had reached just under 9000 cycles. While the FOT trending of the systems shows no signs of motor degradation or slippage, the concern for staying on Profile G indefinitely was that the result in a large number of cycles could potenti ally lead to a solar array drive failure before the fuel projected mission end of life [5]. The concern was prompted from lessons learned from GPM’s predec essor TRMM. Returning to the G Profile to increase drag was resulting in a faster increase in the number of cycles which increased the chance of solar array drive failure before fuel ran out. The rate can be observed in Figure 16, where the green trend represents the number of cycles that would have been encountered if cont inuous tracking had been used since launch and the red trend is what was actually used when the FOT switched between G and K with G being the portion with higher slope . While the orbit requirements w ere being maintained under Profile G, in addition to the increased number of cycles , Profile G was discovered to cause hardware concerns with the solar arrays at certain beta angles as sunlight was able to reach parts of the SC that had not previously when solar arrays were fixed, resulting to temperature s pikes on SADA [5]. The decision to move away from Profile G was finally aided by the fact that while the increase in drag helped in maintaining the requirements for the first eight months after implementation, future predictions showed that the continual drop in solar activity would ultimately result in repeated violations of the mission requirements starting in early spring of 2017. Therefore, in May of 2017, the tracking profile was switched back to Profile K and continued since. This current profile maximizes the amount of feathering, aimed to reduce the drag caused by the large solar panels during flight by strategically positioning the two wings alternatively depending on the beta angle variations . Feathering the wings minimizes the projected area in G G K K K K G G

Document

AMS_2020.pdf

568 the velocity vector directions and lowers substantially the drag on the spacecraft which reduces the number of propulsion burns required to maintain the SC in the desired altitude . In addition feathering also reduces the solar array drive movement s resulting in less number of cycles and wear on the actuators . As can be seen in Figure 16, the use of Profile K has so far resulted in savings of 14,000 cycles on the –Y actuator and 23,000 cycles on the +Y one. On the other hand the reduction in propulsion maneuvers has already resulted in considerable savi ngs in propulsion fuel to date which will allow GPM to have a longer mission life [5]. Currently the GPM Core SC is expected to be in service until 2035. Figure 17. On orbit Drag for Tracking Profiles G & K The behavior of the actuator is monitored by the Flight Operations Team at the Missions Operations Center located at GSFC . Temperatures, currents, voltages, and othe parameters are tracked and supervised to ensure no anomalies occur, or should they occur, to generate the corresponding corrective action. The performance of the +Y and – Y SADAs has so far been remarkable. As can be seen on Figure 18, Commanded vs. A ctual MAX and MIN differences are just 2 steps for both actuators . This is expecte d since the command steps occur before the rotor motion. S o when the telemetry is sampled determines the difference. A step difference of 6 or great er means that the motor missed commutation steps . The ±2 Min & Max levels are consistent with ground testing and beginning of life which indicates the motors, encoder, and drive electronics are all performing nominally . As shown also on Figure 18, the motor current on both +Y and –Y SADAs has consistently remained at approximately 300 mA . Because the motors are using a voltage drive, the current being co nstant indicates that the motor back electromotive force and the resistance are not changing. This telemetry shows that both SADAs have healthy windings and the magnetics (motor constant) are nominal . The encoder consists of a metal disk with slots t hat block or allow light from a Light Emitting Diode (LED) to hit a phototransistor . The 6 tracks include 3 tracks on the motor rotor side to indicate each step and a once around index, and 3 on the harmonic drive gear output to indicate the output angle at unique locations . The HEMI (Hemisphere) Encoder Track is dark for one- half of the range of motion, and light for the other half. The voltage plots shown on Figure 19 show how the maximum Voltage has remained steady at around 14.5 V for both the +Y and –Y SADAs indicating that when the light is blocked, the phototransistor leakage current is small and unchanged. The minimum voltage is also steady at around 0.2 V showing that the LED is not dimming, and the phototransistor current transfer ratios are nominal . The LED driver current set point has not needed to be changed due to aging or environmental degradation in the optoelectronic parts. Both SADA motors are driven using a single drive card, the Mechanism Control Electronics (MCE), which accepts the spacecraft bus voltage and a 1553 interface. The card sends an analog housekeeping packet updated at 1 Hz that includes the secondary voltage rails, the LED currents for each motor, and several thermistors on the card. The MCE also telemeters detailed status on the motor command steps, positions, and commanded parameters at 10 Hz. The SADAs and the MCE telemetry continue to show consistent performance from beginning of life to the present . As seen on Figure 19, t he –Y/+Y actuators temperature has also remai ned steady at an average of 14 and 11°C respectively, indicating the thermal environment

Document

AMS_2020.pdf

569 is well controlled by the thermal system. This is much better than the 31°C on the –Y TRMM SADA that resulted in the anomaly . Temperatures of the cable wrap are also within the nominal desired values . The maximum Hinge Interface Plate temperature (closest to actuator output flange) is 35°C below the 40°C requirement indicating that the thermal design is working properly , and unlike on the TRMM – Y SADA, the GPM SADA ac tuators have good lubrication of bearings and harmonic drive for proper actuator life. . Figure 18. GPM SADA Performance Plots

Document

AMS_2020.pdf

570 Figure 19. GPM SADA Performance Plots

Document

AMS_2020.pdf

571 Lessons Learned For scenarios where a high actuator bending stiffness is required, the use of SuperduplexTM bearings with a high contact angle and preload mounted with a very tight slip fit can be advantageous . A thorough harmonic drive accelerated life -testing program wi th different combinations of circular splines and flexible splines materials can help select the right combination that provides the desired actuator life. Ensuring proper lubrication of the actuator bearings and harmonic drive is critical to the life of t he SADA. Performing a detailed thermal analysis including all heat sources such as internal power, I2 R Joule heating (including harness going through the quill shaft), environmental loads, etc., is critical to the success of the design. This is because lubricant temperatures need to be within their operating ranges, otherwise critical components such as the harmonic drive wave generator bearing / flex spline / circular spline interface and the outboard motor bearing would prematurely fail . These heating loses can be very significant dumping heat directly into the actuator and potentially increasing lubricant depletion. To this effect, the use of a ‘thermal tube’ in the quill shaft helps conduct the majority of the heat away fr om the actuator . Having a robust thermal design with oversized radiators, actuator output sunshades, and redundant heaters is always beneficial . Testing of different Solar Array tracking profiles that allow maximization of the amount of feathering helps reduce the number of propulsion burns required to maintain the SC in the desired altitude as well as reduce the SADA number of cycles and wear on the actuator . Acknowledgements The succes s of the GPM SADA would have never been possible without the outstanding support of NASA Goddard Space Flight Center’s electro- mechanical, mechanical, and thermal engineering branches ’ personnel and supporting contractors . The authors would like to thank everyone for their remarkable efforts . Special recognition to Sierra Nevada Corporation’s (SNC) Durham, NC facility for being so accommodating and cooperative. References 1. Farley, R. E. and Ngo, S. “Development of the Solar Array Deployment and Drive System for the XTE Spacecraft”, Proceedings of the 29th Aerospace Mechanisms Symposium, NASA Johnson Space Flight Center, May 1995, pp. 268- 282. 2. Habibvand, A., “Super Duplex Bearing vs. Simplex Duplex Bearing Technical Memorandum”, Industrial Tectonics Bearings Corporation, Rancho Dominquez, CA, May 1998. 3. Mobl ey, J. and Parker, J. “Harmonic DriveTM Gear Material Selection and Life Testing”, Proceedings of the 41st Aerospace Mechanisms Symposium, Jet Propulsion Laboratory, May 2012, pp. 39- 52. 4. Ward, J., “GPM SADA Cable Wrap Mechanical Design Details ”, NASA GSFC, Greenbelt, MD, July 2009 . 5. Patano, S. et al, “ GPM Orbital Maintenance Planning and Operations in Low Solar Activity Environment ”, 2018 SpaceOps Conference, Marseille, France, May 2018.

Document

AMS_2020.pdf

573 Two-Axis Antenna Pointing Mechanism Qualification f or Juice Mission Dual -Band Medium Gain Antenna Jorge Vázquez*, Mikel Prieto* , Jon Laguna* and Antonio Gonzalez * Abstract JUICE is an ESA mission travelling to Jupiter to get information from the Jovian environment. The Medium Gain Antenna will be used in the Telemetry, Tracking and Command Subsystem of the satellite and serve as the communication antenna d uring close -to-sun phases. It is a dual band antenna (X and Ka band) with a two- axis high accuracy pointing mechanism, which serves also as a deployment actuator. The antenna is fixe d to the spacecraft panel by means of two hold- downs. In this paper , a subsyste m description and brief summar y of the qualification campaign is presented, along with main lessons learned and test conclusions. Introduction The Medium Gain Antenna Major Assembly (MGAMA) is a medium- gain dual -axis steerable antenna assembly that provi des the main uplink and downlink communications between the JUICE spacecraft and Earth. The MGAMA will work on the following bands: • A 2-way X -Band link • A 2-way Ka -Band link Based on BEPI -Colombo and SolO heritage, the main challenges of the design are due to the nature of the mission. Some of them are: • High temperature range in the Antenna Pointing Mechanism ( APM ), due to the combination of hot environment during cruise phase of the mission (up to +150ºC in the APM) and very cold environment in the science phase ( -90ºC in some parts of the mechanism). • Dedicated output shaft support with a double full duplex bearing and a dedicated flexible coupling configuration to get high torsional stiffness and adequate structural strength without overloading internally the gearhead motor . • High accuracy required, on the order of 0.02 º, with an Inductosyn sensor for fine position feedback in the rotation axis. For the reference acquisition, travel -to-end-stop strategy is implemented in the electronics. • High solar radiation and contamination requirements, with a dedicated thermal protection shield design based on an MLI blanket . • Very stringent heat flux requirement from the antenna to the spacecraft. • Low mass. • Need of 358º angular range for elevation axis. * Sener Aeroespacial, Getxo, Spain jorge.vazquez@aeroespacial.sener Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 2020

Document

AMS_2020.pdf

574 Design Description Main characteristics of the antenna are presented in Table 1. Table 1. Medium Gain Antenna Main characteristics Transmit and R ecieve bands X band and Ka band (dual) Peak Gain (X-band) 28.6 dB i Uplink - 30 dB i Downlink Peak Gain (Ka -band) 39.8 dB i Uplink - 39.3 dB i Downlink RF Power handling 50W (X band) - 20W (Ka Band) Main Reflector diameter 510 mm Antenna type Axially Displaced Ellipse Antenna RF transmission element Waveguide Deployment angle 180º Azimuth pointing range 210º Elevation pointing range 358º Mechanism temperature range -40º +150ºC Accuracy (both axis) 0.02º Sensor type Inductosyn Stepper motor type Permanent Magnet 2-Phase 6 -Pole Power conssumption (per axis) < 5W at ambient temperature Total weight (including mechanism) 27.4 kg The main components of the antenna are presented in Figure 1. Figure 1. Medium Gain Antenna 3D view with main components

Document

AMS_2020.pdf

575 • The ARA (Antenna Reflector Assembly), represented in purple in Figure 1, is in charge of supporting the radiating elements of the antenna, and includes the 510-mm Main Reflector, Sub Reflector, ARA Bracket, Feed and struts for supporting the Sub Reflector. • The Boom assembly is composed of the structural components that grant support of the ARA and connect ARA with the APM, allowing the ARA pointing and deploying. • The APM is responsible for the pointing and deploying of the MGAMA components and contains the only permanent interface with the spacecraft . • The HDRM 1 connects the output of the APM to the spacecraft panel to protect the mechanism from excessive loading when in launch conditions. The ARA is fixed by HDRM2. • The Radiofrequency Chain is composed of two waveguide sections (X -band and Ka -band waveguides) in charge of connecting the feed in the ARA with the rotatory joint in the APM. They are supported to the boom by means of calibrated blades . Pointing Mechanism The pointing mechanism from MGAMA is a dual -axis gimbal providing azimuth and elevation steering capability. The azimuth axis is driven by the gearhead motor (GHM) geared to a rotating bracket supporting the elevation actuator which is linked to the MGAMA boom. Both are based on stepper motors with planetary reducers geared to the corresponding output brackets. An integrated X /Ka dual-band and dual - axis Rotary Joint Assembly (RJA) routes the RF energy through the APM in both transmit and receive directions. The APM components are shown in Figure 2. Figure 2. APM 3D view (left) and APM Protoflight model (right) The actuator is based on a gearhead motor connected to a pinion with an elastic coupling in between to isolate the low speed stage from bending moments and shear forces . The pinion is actuating on the wheel (ouput shaft) where an antibacklash gear has been incorporated to eliminate play between the teeth.

Document

AMS_2020.pdf

576 The pinion, main gear and antibacklash wheel are made of non- magnetic material , with dry lubrication (sputtered MoS 2) and hardening coating on the external surface to increase the pitting resistance. In the inside of the gearhead motor, as it can be sealed and temper ature is conveniently controlled by means of dedicated heaters, wet lubrication has been selected (Braycote 601). This combination of dry and wet lubrication made it possible to get adequate tribological behavior of the mechanism components while avoiding lube loss due to evaporation or wet lubricant issues in very cold conditions. The anti -backlash wheel consists of an outer geared ring which is connected to the center shaft by means of 6 calibrated blades. These blades allow a deformation of 0.4º (0.5 mm at the teeth) between the central and outer part of the wheel . The center shaft is connected to the APM main shaft whereas the outer ring is preloaded to the pinion to get the backlash reduction. In Figure 3 there is a 3D view of the spur gear and the anti -backlash design. Figure 3. Main Gear 3D view The feedback of the position is obtained by means of an inductosyn sensor. The Inductosyn transducer is a position sensor made of two plates with square wave patterns printed on them, which become the primary (excitation in rotor) and secondary (sin/cos patterns in the stator) of an electrical transformer. The phase between the input and output varies in the relative movement, resulting in a 180º phase change for a displacement of a coil angular range (360º/Number of poles). Figure 4. Inductosyn sensor

Document

AMS_2020.pdf

577 Twist Capsule The harness is a complex subsystem in these mechanisms due to the huge amount of lines to be tranferred from the stator to the rotor. For this purpose, in azimuth axis, a goose- neck coil supports a metal foil which goes from the external fixed part to the cable drum (internal rotating part). The foil alternates areas for engaging (meshing) and areas for fixation by dedicated brackets. This configuration uses the transfer of inner drum to external housing by bending and requires some means to guarantee that the harness does not slide with respect to the inner drum when unwrapping from the internal drum to the external one. In order to avoid sliding in the internal drum and get proper alignment in the external support, meshing teeth and slots were implemented in the current twist capsule and support elements. In order to guarantee the initial allocation of the twist capsule foil with respect to the teeth pattern, the gap at the slots on the extreme positions is reduced to match exactly the dimension of the teeth. Several iterations on breadboards were done to refine the concept, and the cable attachment before and after the twist capsule were validated by test . The torque behavior of the twist capsule creates a single direction torque which helps to absorb gaps and backlash. The f inal proposed solution is the one implemented in Solar Orbiter and Bepi -Colombo HGAMA that was successfully qualified. The arrangement of the cables in the twist capsule intends to isolate the noisy signals from sensitive ones. In this sense the inductosyn sensor signals from the azimuth stator disk are separated from the rotor excitation and the high current motor lines. For the elevation twist capsule, only two twisted -shielded pair cables (inductosyn excitation main and redundant) shall travel from stator to rotor. In this twist capsule, spiral configuration is used , with cable configured as a long spiral that is wrapped more in the internal drum or the external one, depending on the shaft position. The total number of turns shal l be selected carefully in order to provide the required motion without undesired cable entrapment . The surfaces of the internal drum in contact with the harness shall be soft and corner -free in order to allow sliding between coils. Figure 5. Harness arrangement (left) and azimuth twist capsule (right) Hold Down and Release M echanism (HDRM) The MGAMA is fixed to the spacecraft panel during launch environment by means of two HDRMs. The HDRM 1 is located close to the APM and the HDRM 2 just below the ARA. In this HDRM the movable plate is merged with the ARA bracket and the fixed support is separated into t hree different pieces. The central

Document

AMS_2020.pdf

578 one only supports the N on-Explosive Actuator (NEA) and the cone, whereas the other two support a pair of spheres (female) each. In both the NEA support and the sphere support there are blades that provide high stiffness in lateral and vertical axis and some flexibility in longitudinal direction (boom axis) to absorb thermoelastic distortions. By integrating the HDRMs with the boom and ARA , a low mass was achieved and considerable reduction of loads transferred to the contact surfaces. The combination of spheres and central cone and proper selection of the distance between these components was done to optimize strength and stiffness of the complete subsystem. Figure 6 present s the HDRMs location in the MGAMA. Figure 6. HDRMs location in the MGAMA Figure 7 shows the release process. When the rod of the NEA is released, the spring raises the bolt up to the point where the rod makes contact with the upper surface of the hole in the central cone, acting as an end stop for the bolt+rod assembly. The rod remains inside this cavity to avoid any clash of the movable part with the fixed assembly during deployment. Figure 7. Release sequence

Document

AMS_2020.pdf

579 Mechanism analysis One of the key analysis for the proper dimensioning of the actuator is the torque budget. For this purpose, the combination of the angular speed and acceleration of the spacecraft together with the motion profile characteristics shall be considered. By the calculation of the angular speed (Eq.1) and acceleration ( Eq.2) of the antenna movable part and also the acceleration of its center of gravity (Eq.3), the overall f orce (Eq.4) and moment (Eq.5) on the APM can be obtained. Figure 8 shows the axis convention system together with a 3D plot of the torque as a function of azimuth and elevation angle. Figure 8. Axis convention system and torque in both axis as a function of AZ -El angles The performance analysis has been executed in M ATLAB/Simulink with the Dynamic Performance Model, which includes a multibody definition of the antenna with SimMechanics and the main non- linearities and disturbances (e.g. friction, twist -capsule or backlash).

Document

AMS_2020.pdf

580 The Dynamic Performance Model is a mathematical model which allows, at least, the following analysis: • Microvibration analysis. • Mathematical description for AOCS compatibility verification. • Multibody dynamic model. • Simulation of MGAMA behavior. Figure 9 shows the structure of main blocks of Dynamic Performance Model . Figure 9. Top -level simulator view In Figure 10 disturbance results in the motion phase are presented. Figure 10. Disturbance results in motion phase APM testing Life Model test campaign For the qualification of the poi nting mechanism, a dedicated Life M odel was manufactured and assembled. This unit was submitted to the following tests , from which only the most significant will be presented: • Holding and unpowered torque test • Accuracy, resolution and repeatability test • Stall test • Vibration test • TVAC cycling and life test • Microv ibration test For the functional test , a dedicated test set -up was used with a magnetic strip sensor and a torque cell (see left picture in Figure 11) . There is another picture of the micro- vibration test set -up on the right side of Figure 11. 52 54 56 58-0.6-0.4-0.200.20.4 time [s]Torque [N·m] 52 54 56 58-0.500.5 time [s]Force [N]

Document

AMS_2020.pdf

581 . Figure 11. Test s et-up for APM functional (left) and micro -vibration (right) test The accuracy, resolution and repeatibility test was performed in a differential sequence, making small movements , which were always a discrete number of steps (on the order of 0.1º) and monitoring the commanded and the real movement. T he error curve obtained show very good values with accuracy o n the order of 0.02º and some wave error across the angular range. Figure 12 shows the accuracy test results for Juice L ife Model . • The orange curve plots the difference between the magnetic sensor and the inductosyn. • The blue curve plots the difference between two consecutive samples measured by the inductosyn • The red curve represents the difference between two consecutive samples measured by the magnetic sensor. Figure 12. APM L ife Model accuracy curve before life test During the micro- viration test the exported torques and forces were measured to characterize the mechanism (Figure 13) .

Document

AMS_2020.pdf

582 Figure 13. APM L ife Model exported torques and forces during microvibration test The vibration test (sine and random) was performed without significant issues, with the exception of some connector loose bolts which were not properly tightened (Figure 14) . Figure 14. APM L ife Model Vibration test set -up for Y and Z axis Life test and thermal cycling was performed in a TVAC chamber between the temperature range -50ºC to +150ºC . The movement of the mechanism was monitored during the entire test. The lifecycle is composed of 2600 short -amplitude movements (1º) and 750 long- amplitude move ments (360º), making a total of 2.2 million revs at the motor level. Figure 15 shows the temperature profile and the test set-up in the vacuum chamber.

Document

AMS_2020.pdf

583 Figure 15. APM Life Model Life test temperature profile (top) and test set -up (bottom) After the life test, the accuracy, repeatibility and resolution test was performed again to check that there had not been any significant degradation in the performance of the pointing function of the actuator (Figure 16) . Figure 16. APM Life Model accuracy curve after life test

Document

AMS_2020.pdf

584 The final status of the internal components of the APM was very good in general terms. Even if s ome local depletion of MoS 2 was found in some areas , the general status of the teeth flanges was acceptable with no pitting or surface degradation evidence . Figure 17 shows a picture of the final status of the pinion, antibacklash wheel and main gear teeth. Figure 17. Gears status after life test Protoflight model test campaign After the succesful test campaign with the Life Model , the Protoflight model assembly, integration and test was performed. As usual, functional , vibration and TVAC test s were performed and there were no significant issues. Some pictures are shown in Figure 18. Figure 18. APM Protoflight model vibration test (left) and TVAC test (right) Another key point in the APM test campaign was the magneti c characterization. For this test, a special facility located at Airbus Friedrichshafen was used. Magnetic deperm ing was performed with a magnetic moment measurement before and after to confirm that the reduction was effective. After this process, a total magnetic moment of 136 mA/m2 (combined value in the 3 ax es X Y Z) was obtained. Magnetic susceptibility was also mea sured. Figure 19 shows a picture of this test.

Document

AMS_2020.pdf

585 Figure 19. APM Protoflight model magnetic characterization test The MLI installation on the APM was done after magnetic testing . The areas in which the rotor and the stator MLI are close to each other shall be covered with extreme precautions to avoid any clash or mechanical blockage. The use of li ps or ribs to separate physi cally the MLI blankets is strongly recommended in these areas. Other point s to keep strictly under control are the gaps between MLI blankets, especially if they can clash or get in contact with the spacecraft panels or other subsystems. Figure 20 shows a photograph of the MLI installed on the APM Protoflight model . Figure 20. APM Protoflight model covered with MLI The installation of the mechanism in the final antenna assembly requires a proper fitting of the mechanical interfaces with respect to the structural parts (boom with APM output ) and also the radio frequency flanges from the waveguides. Once performed, the following tests were performed at antenna level: • Functional tests, with deployment of the mechanism at clean room conditions • Radio frequency tests at antenna level • Vibration tests (sine) • Acoustic test • EMC tests • Radio frequency tests See in Figure 21 the deployment (left) and the acoustic test (right) at antenna level.

Document

AMS_2020.pdf

586 Figure 21. MGAMA Functional (left) and Acoustic Test (right) Conclusions During the development and qualification phas e of the JUICE MGAMA, a two -axis pointing mechanism for high accuracy applications and severe thermal environments has been developed. A qualification test campaign has been performed, from which the following conclusions have been extracted: • Combination of wet lubrication in the areas t hat can be sealed and dry lubrication in the areas that cannot is a valid approach, if the temperatures on the oil can be maintained in the temperature range between -40ºC and +150ºC . • For proper behavior of t he dry lubrication it is very important to contr ol the degradation of the MoS 2 layer by avoiding excessive pitting stress on it. Better to avoid this lubrication in gears with very high torques in relation to the pitch of the teeth. • The MLI on mechanisms can help to control the temperatures and reduce the heat flux onto the spacecraft, but the MLI definition and installation shall be done carefully with a detailed design of the blanket’s configuration. Improvisation during mechanism MLI installation is not allowed. • For mechanisms with magnetic cleanliness requirements, deperm is mandatory. However, some misbalance of the magnets of the motors and the stainless steel present in some components may increase the minimum magnetic moment above the required value if this is below 150 mA/m2. • When the mechanism size and weig ht is comparable to the panel to which it is attached, the acoustic test in replacement of random is the best approach to avoid overstressing the components in a non- realist ic manner. References 1. P. Campo, A. Barrio “Development Of A High Temperature Antenna Pointing Mechanisms For Bepi - Colombo Planetary Orbiter”. ESMATS 2013. 2. P. Campo, A. Barrio, F. Martin “Testing of Bepi Colombo Antenna Pointing Mechanism”. ESMATS 2015. 3. J. Vázquez, I. Pinto, I. Gabiola, I. Ibargoyen, F. Martín “Antenna Pointing Mechanisms for Solar Orbiter High and Medium Gain Antennas”. ESMATS 2015

Document

AMS_2020.pdf