Datasets:

ai-aerospace

/

ams_data_full_2000-2020

like 0

293 Figure 8. Results from measurements of oil mobility after wear in the high vacuum tribometer Figure 9. Optical micrographs of the ball and coupon surfaces after wear in the tribometer, showing patterns of lubricant wear product deposition Film Stability and Wetting Tests Resupply of oil to the active tribological contacts also depends on the stability and wettability of the fluid on the substrate. These, too, can depend on composition and film thickness . Thin film stability tests have been done to quantitatively evaluate wetting versus dewetting behavior by monitoring static evolution of a thi n fluid film (h ~3 µm) deposited on a surface (e.g. 440C, 52100, SiOx). This is highly instructive for multi - component systems, such as formulated Pennzane oils, where additives can have an interfacial preference and modify the interfacial landscape (γsl, the surface tension at the solid/liquid interface) . Each wetting test was conducted on two oxygen plasma cleaned 440C steel coupons [10]. Each test fluid was drop cast onto a coupon at room temperature. The fluids were cast from 100 µl of heptane solution that was diluted to 2.5% by volume. The drop cast technique preserves the test fluid stoichiometry and enhances diffusion for a uniform film that is close to its thermodynamically equilibrium state. Once applied,

Document

AMS_2020.pdf

294 the coupon is covered with a petri dish to allow time for stabilization. After 15 minutes, the dish is removed to allow the solvent to evaporate. Two photos are then acquired every 15 minutes with the coupon at room temperature. The coupon is then placed on a hot plate at 40° C for 15 minutes and t hen photographed again. Figure 10 shows time dependent results for three different lubricant component samples. The top row shows thin film stability of a phosphate additive commonly used in aerospace mechanisms. After 15 minutes, there are numerous loc alized regions of dewetting across the surface. This fluid does not maintain a stable fluid on clean 440C steel. The second row shows the time sequence of a thin film of unformulated MAC oil. Dewetting is initiated at the edges of the coupon at room temper ature, and the film rapidly retracts at 40 °C. A thin film of MAC oil, formulated with the phosphate additive, is shown in the third row. Here, we find that the film is stable at all temperatures that were tested. It is evident that the additives are essent ial to maintain a stable film of oil in the laboratory environment. Figure 10. Photographs of steel coupons during oil wettability tests Oil Meniscus Evolution To help prolong the life of moving mechanisms, the l essons learned regarding mobility of lubricants must be ultimately applied to models of the supply rate of oil to the ball/race interface. The ultimate objective is to devise and test strategies for sustaining oil and promoting oil resupply at the critical ball /race interfaces. These efforts are guided by a theoretical model of oil uptake from the surrounding surfaces. As oil is drawn into the meniscus via surface tension and capillary forces, it fills the volume defined by a cylindrically symmetric annulus, with measurable base (b), waist, and cap ( a) radii. A time interval optical microscopic imaging system was constructed to record images of a physical contact while the oil is flooding the meniscus .[8, 12] This system enables dynamic volume measurements over long periods of time. A screenshot of the meniscus analysis system is shown in F igure 11. Here, the meniscus profile, meniscus

Document

AMS_2020.pdf

295 volume and other elements describing the meniscus geometry are extracted and measured from optical images. Figure 11. Screenshot showing the imaging window in a home- made application to measure the uptake of oil in the meniscus between a ball and flat. An example of the evolution of these parameters during oil uptake in the meniscus is shown in F igure 12. In this case, the formulated MAC oil was cast as a thin film of 2.85 µm thickness on a clean 440C steel substrate at 25 °C. A 440C ball was then brought into contact with the film at time t = 0, and the cap radius (a), meniscus height (h), and meniscus volume were monitored with time using the automatic imaging system. In this case , the meniscus growth was monitored for the first 180 hr and was only parti ally filled at that point. Each parameter was normalized to its value at 17 0 hr, so that they can be plotted together. We find continuous growth in volume that starts immediately upon contact, and is taken up at a rate that is controlled by thin film flow on the substrate and the changing capillary pressure as the ball/race interface floods. Figure 13 shows the meniscus volume as a function of time, using a more viscous gyroscope oil. This series of tests was run until the volume saturated after approximately one day of uptake from a thin film. The data are compared against the results of a theoretical model which will be described elsewhere.

Document

AMS_2020.pdf

296 Figure 12. Three different geometric parameters of an MAC oil meniscus are shown as the capillary between a ball and flat is flooded with oil from a thin film on the flat Figure 13. Time evolution of a ball/flat oil meniscus flooded by a thin film on the flat substrate Effects of Composition on Oil Film Thickness Changes in oil composition and mobility can ultimately affect the thickness and starvation of a protective oil film in the rolling contacts of high- speed bearings. The results above show the effective increase in viscosity of oil in the vicinity of a tribological contact due to polymeri zation. These results show the effects of changes in mobility due to deposition of lubricant degradation products, and illustrates that these products remain in the active tribological contacts. This would alter the molecular weight distribution as the mechanism ages, skewing it to heavier molecules and higher effective viscosity. To study the potential effects of aging on lubricant film thickness in spacecraft mechanisms, we have conducted EHD film thickness measurements on operating angular contact bearings using lubricants of similar chemical constituents and various molecular weights.

Document

AMS_2020.pdf

297 Film thickness measurements were performed using a device that has been described elsewhere [13-16]. Two angular contact bearings are mounted in a spring preloaded DF arrangement. As the EHD film thickness increases with speed, the resulting deflection of the outer rings is measured with capacitance proximity sensors. This displacement is simultaneously determined from changes in the dynamic preload using a load cell placed in series with the bearing preload. Figure 14 shows a schematic representation of the bearing deflection resulting from film thickness growth. At low speed, the balls are in intimate contact with the raceway surfaces. At high speed, the gr owth of the EHD film alters the bearing geometry, displacing the outer race to the right, and reducing the contact angle from b 1 to b 2. The change in film thickness is inferred from the outer ring displacement using a geometric model that represents the fi lm thickness as an effective change in the ball diameter from d 1 to d 2. The change in contact angle is determined by measuring changes in the ball group frequency with respect to the shaft speed. Figure 14. Cross section of an angular contact bearing at low and high speeds, showing effects of increasing EHD film thickness, modeled as an effective increase in ball diameter Tests were conducted using poly alpha olefin samples of various viscosity. These lubricants were chosen due to their relevance to aerospace applications, their availability with different viscosities, and their simple composition compared to refined mineral oils. The oil samples were not formulated with additives. A small amount of oil was appli ed to bearings by first impregnating the retainers, and then centrifuging the retainer to remove bulk oil from the surface. The metal parts were coated with a thin film by solvent casting from a 10% solution of oil in heptane solvent. After running- in the lubricant at 3000 rpm for 500 hours at room temperature, film thickness vs speed profiles were collected. These experiments were conducted by operating the bearings at a chosen speed, and then quickly stopping the shaft to collapse the film. The resulti ng rapid change in outer ring position was recorded and converted to oil film thickness using our geometric relationship. By measuring an instantaneous change in position, we avoid effects of thermal relaxation on the bearing structure that would otherwise overwhelm and obscure changes in the oil film thickness. This process was repeated after sequential increases in shaft speed from 60 to 3000 rpm. An example of our results is shown in Figure 15, where EHD film thickness of two different PAO oils is plotted against speed from 60 to 3000 rpm. Orange circles are results from tests with a low viscosity PAO (10 cS at 100° C), and blue circles are from a more viscous oil (40 cS at 100° C). Tests were done with the bearings at 30° C. We find that the PAO40 fil m thickness increases much more rapidly with speed than the PAO10, as expected by theoretical estimates based on the Hamrock Dowson film thickness model. [17, 18] The more interesting observation, however, is the deviation from this model at high speeds due to kinematic starvation of the rolling contact. [19] We find that in this case the onset of starvation occurs at an approximately 600 rpm for the PAO40 and 900 for the PAO10. The difference i s likely due to the reduced

Document

AMS_2020.pdf

298 mobility of the more viscous oil, impeding the reflow of oil from the periphery of the contact back into the inlet zone. These results show the degree to which film thickness depends on viscosity. As lubricant deteriorates in an aging satellite mechanism, we may expect similar increases in film thickness. However, the experiments shown here were performed after only 500 hours of operation. While this is enough time for run- in of the lubricant and stabilization of the film thickness, it is not sufficiently long to reach the more severely starved conditions of a late- life satellite mechanism. For that reason, we suspect these film thicknesses are not representative of such conditions. Future experiments will explor e the effects of lubricant depletion. Figure 15. Dependence of EHD film thickness on speed, for two lubricants with similar chemistry and molecular architecture, but different molecular weight and viscosity. Circles are experimental measurements; solid lines are fits to Hamrock -Dowson model Conclusions Lubrication is an essential component of any moving mechanical assembly. In most space mechanisms, where the lubricant cannot be easily replenished, degradation and loss of the oil ultimately leads to changes in performance and an end to component life. Our goals have been to develop a better understanding of how to manage those changes in performance and to delay the end of life. The experimental techniques and the selected results shown here demonstrate our approach to achieving these goals. Among the lessons learned from this study are that: 1) Spacecraft component life often depends on management of thin films in an evolving environment, 2) Oil mobility decreases with film thickness, 3) Oil mobili ty decreases as it degrades in a tribological contact, 3) new techniques can detect subtle changes in lubricant composition, 4) these changes can affect wettability, mobility, and resupply of oil to critical contacts, and 5) impeded resupply can reduce the thickness and stability of a protective film between two surfaces. 0510152025303540 0 500 1000 1500 2000 2500 3000EHD Film Thickness ( µin) Rotational Speed (rpm)Comparison of PAO10 with PAO40 Oil PAO40 Experiment PAO10 Experiment PAO40 Theory PAO10 Theory

Document

AMS_2020.pdf

299 References 1. Palmgren, A. and B. Snare, Influence of Load and Motion on the Lubrication and Wear of Rolling Elements. Inst. Of Mech. Eng., 1957. 79 : p. 454- 458. 2. Helt, J.M., Density, Dynamic and Kinematic Viscosity of Bearing Lubricating Oils and Additives . 2019, The Aerospace Corporation: El Segundo, CA. 3. Messaâdi, A., et al., A New Equation Relating the Viscosity Arrhenius Temperature and the Activation Energy for Some Newtonian Classical Solvents. Journal of Chemistry, 2015. 2015(3): p. 12 pages. 4. Deryagin, B.V. and V.V. Karasev, The Study of the Boundary Viscosity of Organic Liquids by the Blow -off Method. Russian Chemical Reviews, 1988. 57(7): p. 634- 647. 5. Scarpulla, M.A., C.M. Mate, and M.D. Carter, Air shear driven flow of thin perfluoropolyether polymer films. The Journal of Chemical Physics, 2003. 118(7): p. 3368- 3375. 6. Berendsen, C.W.J., et al., Rupture of Thin Liquid Films Induced by Impinging Air -Jets. Langmuir, 2012. 28(26): p. 9977- 9985. 7. Helt, J.M., High -Vacuum Tribometry Tests of Fluid Lubricants on Bearing Steels . 2013, The Aerospace Corporation: El Segundo, CA. 8. Helt, J.M. and P.F. Frantz, Investigating the Reflow of Gyr oscope Lubricant SRS160 on Bearing Surfaces: Analysis Methods and Techniques . 2013, The Aerospace Corporation: El Segundo, CA. 9. Helt, J.M., High vacuum tribometry tests of fluid lubricants on bearing steels , in ACS National Meeting . 2014: San Francisco C A. 10. Helt, J.M., Procedure for Separating Pennzane Oil from Formulated Oils and Rheolube Based Greases . 2016, The Aerospace Corporation: El Segundo, CA. 11. Helt, J.M. and J.J. Kirsch, High Vacuum Tribometer for Fluid Lubricated Bearing Material Testing: Part 1 – Instrument Design. 2010, The Aerospace Corporation: El Segundo, CA. 12. Didziulis, S., J. Helt, and P. Frantz, Laboratory Studies of Spacecraft Fluid Lubricants , in ACS National Meeting . 2019: San Diego CA. 13. Ward, P.C., A.R. Leveille, and P.P. Frantz, Measuring the EHD Film Thickness in a Rotating Ball Bearing, in Proceedings of the 39th Aerospace Mechanisms Symposium. 2008: NASA Marshall Space Flight Center. 14. Helt, J.M., EHD Bearing Test Unit: Final Design . 2016, The Aerospace Corporation: El Segundo, CA. 15. Helt, J.M., EHD Bearing Test Unit: Setup & Operation Overview . 2017, The Aerospace Corporation: El Segundo, CA. 16. Helt, J.M., EHD Bearing Test Unit: Automated Test and Analysis Framework . 2018, The Aerospace Corporation: El Segundo, C A. 17. Dowson, D. and G.R. Higginson, Elasto -hydrodynamic lubrication the fundamentals of roller and gear lubrication. 1966. 18. Hamrock, B.J. and D. Dowson, Isothermal Elastohydrodynamic Lubrication of Point Contacts: Part III —Fully Flooded Results. Journ al of Tribology, 1977. 99(2): p. 264- 275. 19. Coy, J.J. and E.V. Zaretsky, Some Limitations in Applying Classical EHD Film Thickness Formulas to a High- Speed Bearing. Journal of Tribology, 1981. 103(2): p. 295- 301.

Document

AMS_2020.pdf

301 Efficacy of Lead Naphthenate for Wear Protection in High Vacuum Space Mechanisms Jason T. Galary* Abstract The purpose of this research is to investigate the efficacy of lead naphthenate as a wear additive in a multi - alkylated cyclopentane (MAC) fluid for use in high vacuum space mechanism applications. The use of lead naphthenate in MAC lubricants has a spacef light history of over thirty years. However, despite the history of use for this additive in a variety of rolling and sliding applications, little is known or understood about the tribochemical process by which these additives function. This research looks at the performance of this additive in simulated contact tests using SRV (Sliding in Atmospheric and Vacuum Conditions ), as well as a scuffing test performed in a mixed rolling/sliding contact. In addition, application simulation tests are being performed including high vacuum spiral orbit tribometry (SOT) and vacuum angular contact ball bearing testing. This test program evaluates the additive in both a mixed film and boundary lubrication contact under vacuum and atmospheric conditions to better understand how the additive functions on a metal surface. This report of the test programs current progress will include high vacuum bearing life testing, boundary/mixed lubrication scuffing wear, outgassing, and SOT. The results of this work will help the design engineer understand how materials, including lubricants, play a critical role in the performance and life of space mechanisms in demanding high vacuum environments. A greater understanding of the relationship between lead content and tribological performance will be developed along with further understanding of the tribochemical degra dation process. Data gathered from wear testing and application simulation work will provide mechanism design engineers with a better understanding of the tribological performance of this lubricant additive . Introduction and Background Lead naphthenate has been used heavily as an anti -wear and extreme pressure additive in multi alkylated cyclopentane lubricants for high vacuum space mechanisms. The additive consists of a centralized lead ion that is bonded with the oxygen atoms of two carboxylate groups each attached to naphthenate aromatic rings . The naphthenate aromatic hydrocarbon rings provide solubility of the lead naphthenate in different hydrocarbon oils , but the way the additive reacts with steel and protects against wear in various lubrication regimes is still not fully understood. The historical use of lead naphthenate originates from industrial gears and bearings where it was used an extreme pressure additive over fifty years ago. Over the last thirty years, it has gained a lot of spaceflight history although there are still many questions regarding its efficacy , primarily how does it function in various lubrication regimes, the effects of different metallurgy, and how environmental pressure effect its performance. With the increasing number of space mechanisms being developed and launched as well as the increasing length of mission time required, it is critical to have a robust design with high reli ability. To improve the reliability, it is necessary to have long lives for all of the components in the design. This will require that the lubricants used for space mechanisms must also improve. The use of lead naphthenate additive in multi alkylated cycl opentane fluids is a proven additive package with spaceflight history but in order to develop * Nye Lubricants, Inc., Fairhaven, MA Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 2020

Document

AMS_2020.pdf

302 and advance the lubrication technology for high vacuum space mechanisms, additional understanding of its tribological performance is needed to facilitate innovations . When considering how to research the efficacy of lead naphthenate and how it functions as a tribofilm, it is important to look at various methodologies to characterize its performance. In the beginning stages, this is easiest performed by using simulated contact testing. In this respect, an SRV makes the best choice as it is very flexible regarding the contact mechanics and envi ronmental conditions. The availability of SOT testing is somewhat limited but a Mini Traction Machine (MTM) can be used in a manner to simulate the same type of application conditions with a ball running in an orbital pattern on a disc with a mixture of rolling/sliding. While the MTM cannot currently run in a vacuum condition, it can be used to understand the fundamental complexities of lead naphthenate operating in a rolling/sliding mixed application. When this is combined with the vacuum/atmospheric testi ng on the SRV, the performance of the lead naphthenate can be made clearer. In more recent research done on lead naphthenate using the SRV and other sliding tribometers, results have typically been inconclusive regarding the efficacy of this additive to protect a lubricated contact. The standard SRV testing includes conditions that could be inappropriat e for evaluating many anti -wear additives including lead naphthenate. This includes the sliding speed which would create a thicker film preventing the additive function from being studied. The contact stress is also much higher in standard ASTM tests which will influence how the tribofilms are created and in the case of lead naphthenate, the higher contact stresses will create additional wear which will react and consume available lead naphthenate making it unavailable to create a protective tribofilm. From the authors previous resear ch [7], the following was determined: In oscillatory testing, Lead Naphthenate had better anti -wear performance at lower temperatures (~20°C) and increasing the amount of lead directly reduced the wear rate. Under high vacuum conditions, the samples with lead naphthenate offered twice the wear protection with 440C performing the best. At higher temperatures (75°C) the wear rate was almost double for the samples with a higher content of lead across all experiments. The 75°C testing in pure oscillatory sliding on 52100 steel, showed that samples with lead naphthenate in an atmospheric environment performed worse as th e concentration of lead increased. However, when in a vacuum environment, as the concentration increased the wear rate decreased. In 440C testing, samples with lead offered up to twice the wear protection over the neat samples in a vacuum environment. In mixed rolling/sliding experiments, Lead Naphthenate had better anti -wear performance at high temperatures and increasing the amount of lead directly reduced the wear rate. At lower temperatures (50°C) in the counter - rotation wear test, lead naphthenate offered three times the wear protection when in a 3% concentration and eleven times the protection for 5% lead compared to the neat MAC fluid. At higher temperatures (150°C) in the counter -rotation wear test, lead naphthenate offered five times the wear protec tion when in a 3% concentration and twenty times the protection for 5% lead compared to the neat MAC fluid. There appears to be is a transition point between 3% and 5% lead naphthenate where the available lead can react with both the surface and worn metal to create a strong lead anti -wear tribofilm. This tribofilm that is created is between 2- 4µm thick and while it will cause an overall increase in friction at the surface, the wear of the contact is greatly reduced. Previous Lessons Learned from AMS 2018: - The additive function of lead naphthenate is a combination of physical absorption through rubbing or pressure and chemisorption. - Depending on the mechanics in the contact (sliding versus rolling), the effect of temperature had a significant influence . This appears to indicate that higher concentrations of lead would be required for more severe applications involving pure sliding and/or high temperatures as the lead is consumed faster through reaction with surface layer steel oxides and sublayers. - In rolling and mixed contacts, lead naphthenate creates a strong tribofilm on the steel surface that aids in protecting from wear but at the same time will increase the friction in the contact. - The formation and durability of lead naphthenate tribofilm is depende nt on the environment with higher performance coming under vacuum conditions. It is also believed the lack of oxygen

Document

AMS_2020.pdf

303 promotes this life due to the lack of oxide formation on the steel and degradation of the lead naphthenate. - In general, an increase in lead content will decrease the wear rate. Figure 1: Mixed Film Oscillatory Wear Results from AMS 2018 Scope of Work The following work was covered in this research. • Fluid Lubricant Evaluation: The performance of five multi- alkylated cyclopentane (MAC) fluid lubricants were performed included the heritage material Synthetic Oil 2001A, two lead naphthenate versions (2001- 3PB and 2001- 5PB), and comparable versions that utilize triphenyl phosphate (2001B) and tricresyl phosphate (2001T) . • Grease Lubricant Evaluation: The grease counterparts to the materials in the fluid evaluation (Rheolube 2000F , 2004, 2000- 5PB, 2000B, and 2000T ) were tested under a vacuum bearing configur ation. Testing Apparatus Three different tribological test methods were used in this research. One of the methods is run in pure sliding under both atmospheric and vacuum conditions (5 x 10-5 Torr Min) while the other two methods utilize a mixed rolling/sliding contact in atmospheric conditions for both mixed film and boundary lubrication. The friction/traction properties along with the wear rates were measured for two multiply alkylated cyclopentane (MAC) hydrocarbon oils formulated with 3% and 5% lead naphthenate (2001- 3PB and 2001- 5PB) along with unformulated MAC oil (2001A). The lead naphthenate used in these samples was vacuum treated prior to formulation to improve vacuum outgassing characteristics and make it suitable for a space mechanism lubricant. Spiral Orbit Tribometer The Spiral Orbit Tribometer (SOT) is a single contact simulation that replicates Angular Contact Ball Bearings in a semi -starved lubrication mode while under vacuum. It is simply a ball held between two parallel plates which essentially makes it a thrust bearing. The SOT utilizes two flat, concentric discs between which a ½” lubricated ball rolls and pivots during test operation. A very small amount of lubricant, approximately 50μg, is applied to the ball to ensure the system operates in the boundary lubrication regime, providing a fast and efficient screening method of lubricant performance. The upper disc applies a load which can be varied depending on the desired Hertzian Contact Stress (material, load, and bal l diameter dependent). During test operation, the lack of a ball retainer allows the ball to orbit in an opening spiral pattern whose orbit radius continuously increases throughout the course of one revolution of the bottom sample disc. Since

Document

AMS_2020.pdf

304 there is no retainer, if this orbit was left unchecked, it would result in the ball orbiting out of the two parallel sample discs, causing the two plates to contact each other. Fortunately, as the ball reaches the completion of each revolution, it is pushed back into its original orbit radius by contacting what is known as the guide plate. The test is initiated with the sample ball touching the guide plate, which sets the initial orbit radius, and is adjustable. The guide plate is oriented in a position such that it contacts the center of the ball on each revolution. Attached to the guide plate is a charge amplifier force transducer. This force transducer measures the reactionary force required for the ball to be scrubbed back into its initial orbit diameter. This location throughout the orbit is identified as the scrub region. As the lubricant becomes consumed or degrades, this force will naturally increase. Figure 2. Spiral Orbit Tribometer mechanisms, Image courtesy of Spiralab Vacuum Bearing Testing : The apparatus used for these experiments tests a single angular contact ball bearing. The system operates at a vacuum level better than 5.0 x 10- 7 Pa, from 1 to 500 R PM, up to 200°C, and loads to 450 N. The system uses dead weight loading and either a heat lamp or band heater. The system also measures cross bearing electrical resistance, which is used to monitor the operating regime. Bearing torque, load, chamber press ure, and cross bearing resistance are recorded using a data acquisition system. The system uses a single SKF 7204 1219 ( 52100 steel) angular contact bearing. The bearing has an outside diameter of 47 mm, a bore of 20 mm, eleven 12.7 mm balls and a steel retainer. The bearing is mounted in a fixture that holds the outer race and rotates the inner race. Temperature information is gathered from a thermocouple mounted just below the inner race. Wear Testing: Three different wear rates were calculated including a mixed film in pure sliding, mixed film (rolling/sliding), and boundary (rolling/sliding). A normalized wear rate for all these tests were calculated to compare materials tested. The wear rate will indicate the volume of wear (µm3) over a distance traveled (in millimeters) which will normalize the data in the case of premature failures. For the pure sliding test, the ASTM D-5707 SRV Coefficient of Friction test was used. For both the mixed film and boundary testing in a rolling/sliding contact, a custom experimental method using a Mini Traction Machine (MTM) (as show in Figure 3) was used [1][2][4][7][ 13]. In the MTM, the ball and disc are driven independently which allows any combination of rolling and sliding. The measurement of friction force is done through a load cell that is attached to the bearing housing of the ball motor shaft.

Document

AMS_2020.pdf

305 Figure 3. Schematic of Mini Traction Machine, image courtesy of PCS Instruments The experimental method using the MTM utilized a ball -on-disc configuration. In the MTM, the ball and disc are driven independently which allows any combination of rolling and sliding. The measurement of friction force is done through a load cell that is attached to the bearing housing of the ball motor shaft. In previous work by the author and when considering other scuffing tests, it was noted that most simulations use a load stage progression [5,6,13] . This can be seen in the FZG (Forschungsstelle fur Zahnrader und Getriebebau) , 4-Ball EP (Extreme Pressure) , SRV, OK Load Test, and Timken tests . The testing in this work is done using progressive speed as opposed to progressive load stages . On benefit of this approach is that the higher sliding speeds will allow for a more aggressive wear rate to help differentiate the efficacy of the additives . However, the primary reason for using the progressive speed approach is the fact that in a test w here the contact stress increases at every stage, the size of contact patch will also increase at every stage. This leads to fresh nascent metal being exposed at every stage. This new area of contact has not yet developed a tribofilm when it comes into contact , so it is more likely to have aggressive wear. Therefore, tests run in a progressive load methodology will typically have failures at the step increase and have lower repeatability. The experimental testing methodology is as follows [1,7]: The testing apparatus is assembled and filled with oil for the experiment. The temperature of the oil is then heated to 150oC while the ball and disc are rotated at a slow speed while not in contact. This continues for 30 minutes to allow for any chemical absorption of the additives on the surfaces. After this, there is a 10- minute run- in period with a Hertzian contact stress of 1.25 GPa and a Sliding to Rolling Ratio ( SRR ) of 1. Once the run- in has completed, a progressive speed test starts with running stages for 1 min ute and rest stages for 30 seconds. The SRR is varied at each stage in order to maintain the entrainment speed for each stage but increase the sliding speed at the contact. The test continues until either all stages (51) are completed (maximum speed for MT M) or scuffing wear and seizure occurs. Samples The test plates for the SOT were machined from 440C stainless steel with a surface roughness of <0.05 microns. Balls used were 12.7- mm or 7.14- mm 440C stainless steel depending on the contact stress requirements. For the vacuum bearing tests , SKF 7204 Angular contact ball bearings made from 52100 steel were used. In the SRV testing, AISI 52100 steel balls and discs were used with a Young’s Modulus of 210 GPa and a Poisson ratio of 0.30. The balls had a 10- mm diameter with a roughness (Ra) of 25 nm and a Rockwell hardness of 62 HRC. The 24mm discs are vacuum arc re- melted and had a hardness of 58 HRC with a lapped surface that has a roughness (Rz) of 500 nm. All the tests were run in duplicate and with a maximum Hertzian contact stress of 2.12 GPa and a sliding speed of 300 mm/s. The experimental MTM test used AISI 52100 steel balls and discs were with a Young’s Modulus of 210 GPa and a Poisson ratio of 0.30. The balls had a 19.05- mm diameter with a roughness (Rq) of 10 nm and a Rockwell hardness

Document

AMS_2020.pdf

306 of 62.5. The discs had a hardness of 60.5 and a roughness (Rq) of 11 nm. All the tests were run in duplicate and with a maximum Hertzian contact stress of 1.25 GPa. Procedures Sample Preparation All test specimens are ultrasonically cleaned in heptane followed by acetone. For the SOT testing, the fluid lubricants were plated via the preparation of a solution of lubricant diluted into an appropriate solvent. This solution was then applied directly to a rotating ball and the solvent allowed to evaporate from the surface. This left the fluid lubricant on the ball’s surface. The lubricant plating method allows for the application of very smal l amounts of lubricant. The application of grease was done by rubbing the ball between cleanroom grade polyethelye sheets until and even coating of lubricant was applied. The target amount of applied lubricant for both fluids and greases were 50 μg. The contact area for the SRV was coated with 5 ml of lubricant while the MTM specimens were fully flooded. Results and Discussion SRV (Sliding in Atmospheric and Vacuum Conditions) The results in Tables 1- 3 and Figure 4 are from the experiment on the MAC with 0%, 3%, and 5% lead naphthenate when tested in a boundary lubricating regime , at 20°C on the SRV under both atmospheric and vacuum conditions. These plots illustrate the wear rate of each material which is determined by the total wear volume (µm3) per millimeter traveled in the test. The wear volumes were measured using an Ametek 3D Optical Profilometer. Table 1: Wear Performance for MAC fluid with 0% Lead at 20°C Table 2: Wear Performance for MAC fluid with 3% Lead at 20°C Table 3: Wear Performance for MAC fluid with 5% Lead at 20°C Material Temp ( °C)Environment Specimen Avg Wear Scar (mm2)Disc Wear Volume (μm3)Wear Rate (μm3/mm) 0% Lead 20 Atmosphere 440C 1.302 599,972 0.28 0% Lead 20 Vacuum 440C 1.294 518,400 0.24 0% Lead 20 Atmosphere 52100 1.499 481,018 0.22 0% Lead 20 Vacuum 52100 1.420 410,400 0.19 Material Temp ( °C) Environment Specimen Avg Wear Scar (mm2)Disc Wear Volume (μm3)Wear Rate (μm3/mm) 3% Lead 20 Atmosphere 440C 2.409 410,400 0.19 3% Lead 20 Vacuum 440C 1.942 216,000 0.10 3% Lead 20 Atmosphere 52100 1.950 342,356 0.16 3% Lead 20 Vacuum 52100 1.820 259,200 0.12 Material Temp ( °C)Environment Specimen Avg Wear Scar (mm2)Disc Wear Volume (μm3)Wear Rate (μm3/mm) 5% Lead 20 Atmosphere 440C 1.468 291,600 0.14 5% Lead 20 Vacuum 440C 1.259 172,800 0.08 5% Lead 20 Atmosphere 52100 1.389 259,200 0.12 5% Lead 20 Vacuum 52100 1.242 216,000 0.10

Document

AMS_2020.pdf

307 Figure 4: Boundary Lubrication Wear Rate at 20°C In atmospheric conditions, the wear rates for 52100 steel was consistently lower than the 440C. Under vacuum conditions , all of the samples had lower wear rates than the atmospheric tests, the samples with lead had almost half the wear than neat oils, and the amount of lead made a small difference in the wear. When looking at the average wear scar there is no correlation to the disc wear volume of the wear rate. Until recently, most published papers used the average wear scar to compare efficacy of wear additives and performance in testing. Using this measurement simply gives you a dimension of the worn area with no indication of how much material was removed. The results from this study as well as those presented by St. Pierre [9] have illustrated that two-dimensional wear measurements cannot be relied on to understand what is going on in a mechanism or trib ological contact. By using 3D profilometry, a deeper understanding of what is going on can be attained. The results in Tables 4- 6 and Figure 5 are for the multiply alkylated cyclopentane with 0%, 3%, and 5% lead naphthenate when tested in a boundary lubric ating regime, at 75°C on the SRV under both atmospheric and vacuum conditions. Table 4: Wear Performance for MAC fluid with 0% Lead at 75°C Table 5: Wear Performance for MAC fluid with 3% Lead at 75°C Table 6: Wear Performance for MAC fluid with 5% Lead at 75°C Material Temp ( °C)Environment Specimen Avg Wear Scar (mm2) Disc Wear Volume (μm3)Wear Rate (μm3/mm) 0% Lead 75 Atmosphere 440C 1.437 1,080,000 0.50 0% Lead 75 Vacuum 440C 1.786 972,000 0.45 0% Lead 75 Atmosphere 52100 1.513 263,796 0.12 0% Lead 75 Vacuum 52100 1.826 276,480 0.13 Material Temp ( °C)Environment Specimen Avg Wear Scar (mm2) Disc Wear Volume (μm3)Wear Rate (μm3/mm) 3% Lead 75 Atmosphere 440C 1.449 972,000 0.45 3% Lead 75 Vacuum 440C 1.824 691,200 0.32 3% Lead 75 Atmosphere 52100 1.346 622,811 0.29 3% Lead 75 Vacuum 52100 1.236 259,200 0.12 Material Temp ( °C)Environment Specimen Avg Wear Scar (mm2) Disc Wear Volume (μm3) Wear Rate (μm3/mm) 5% Lead 75 Atmosphere 440C 3.003 756000 0.35 5% Lead 75 Vacuum 440C 1.874 388800 0.18 5% Lead 75 Atmosphere 52100 3.335 612415 0.28 5% Lead 75 Vacuum 52100 1.842 162000 0.08

Document

AMS_2020.pdf

308 Figure 5: Boundary Lubrication Wear Rate at 75°C In atmospheric conditions, the wear rate for 52100 steel was consistently lower than the 440C although they trended in opposite directions with the increase in lead content. Under vacuum conditions, the effect for the sample with 0% lead was minimal but the samples with 3% and 5% lead produced considerably lower wear rates for both 440C and 52100. This opposite trend between the 52100 and 440C is believed to be related to the way lead naphthenate interacts with the chemical composition of the steel [10 ]. With 52100, the lead naphthenate reacts with the iron oxide present at the surface layer and created with wear debris. When under vacuum there is less iron oxide formation which a llows the lead naphthenate to provide more wear protection. Regarding the 440C, the lead naphthenate will chemisorb into the chromium layers of the stainless and provide better wear protection as the concentration increases and the environment goes from at mospheric to vacuum. Comparing the wear performance between 20°C and 75°C, all samples had a higher wear rate (2-3X) on 440C at 75°C except for the 5% lead naphthenate sample tested under vacuum. As this SRV testing is a pure sliding test, the lower wear resistance for the 440C at 75°C is tied to the more complex layered structure of the metal. The structure of the 440C would require both chemisorption and physical absorption for the best performance. On the 52100 specimens, the samples with 0% lead had h alf the wear at 75°C, around half the level of wear at a 3% l and 5% loading of lead under the atmospheric tests. In the vacuum tests on 52100, all of the wear rates were comparable. MTM (Rolling/Sliding Contra- Rotation) The results in Figures 6 -7 are for the MAC fluid with 0%, 3%, and 5% lead naphthenate as well as TPP and TCP when tested in a mixed and boundary lubricating regime and under a rolling/sliding configuration on the MTM at 50°C and at 150°C. These plots illustrate the wear rate of each material.

Document

AMS_2020.pdf

309 Figure 6: Mixed Film Scuffing Wear Rate Comparison Figure 7: Boundary Scuffing Wear Rate Comparison From these results, we can see that there is a transition point in the concentration of lead in these MAC fluids and how it react s with the surface metal to form an anti -wear tribofilm. It should be noted, that this phenomenon was not seen in the SRV testi ng. It is believed that the aggressive sliding in the SRV test creates an entirely separate wear mechanism that prevents the lead naphthenate from reacting with the surface and building a strong tribofilm. In the MTM testing, there is a mixed rolling/slidi ng which will promote a tribofilm to be created in a fashion similar to gears and bearings that have a proper run- in process . This would agree and confirms work done by Carre et al [11] where it was found in ball bearing test data that lead naphthenate reacts with metal wear particles to create lead- containing surface coatings.

Document

AMS_2020.pdf

310 Spiral Orbit Tribometry The results in Figure 8 are for the MAC with 0%, 3%, and 5% lead naphthenate as well as TPP and TCP when tested in a pure rolling SOT test under boundary lubrication conditions. The Rheolube 2000F is a polytetrafluoroethylene (PTFE) thickened version of the MAC with 0% lead. In these SOT tests, the TPP outperformed the TCP and lead based additives in relative life. As the SOT does not provide adequate conditions to form tribofilms on the surface of the contacts in the test, the TPP is expected to have performed well due to the decomposition of TPP which creates a multilayered solid film on iron or iron oxide [ 14]. The results for the TCP and the formulations containing lead were slightly surprising but this is expected to tie to the restrictions of the mechanics in this test to create tribofilms and fully simulate a bearings performance. In all of the SOT testing, the Synthetic Oil 2001A (or 2000F grease version) that contained no additives performed the best which is believed to be a combination of the structure of the MAC oil itself, inability for wear additives to create a tribofilms under these test conditions, degradation of the anti -wear additives, and viscous friction. Figure 8: SOT Normalized Lifetimes High Vacuum Bearing Testing The results in Figure 9-10 are for the MAC with 0%, 3%, and 5% lead naphthenate as well as TPP and TCP when tested in angular contact ball bearings under boundary lubrication conditions. The grease versions were also tested with Rheolube 2004 containing 3% lead. All of the greases were thickened with sodium soap. In these bearing tests, the TPP outperformed the TCP similar to the relative life testing performed on the SOT . In bearing tests on both the oil samples and grease versions, the addition of lead naphthenate increased the wear in the bearing which will shorten the lifespan. These results correlate to what was seen in the SOT testing where the lead was not an effective anti -wear additive for these rolling contacts in vacuum.

Document

AMS_2020.pdf

311 Figure 9: Vacuum Bearing Wear Test on Oils Figure 10: Vacuum Bearing Wear Test on Greases In these vacuum bearing tests, the Synthetic Oil 2001A that contained no additives performed the best which is believed to be a combination of the structure of the MAC oil itself, inability for wear additives to create a tribofilms under these test conditions (no mechani cal run -in or chemisorption before the test ), degradation of the anti -wear additives, and viscous friction. Conclusions In previous studies done on lead naphthenate using the SRV and other sliding tribometers, results have typically been inconclusive regarding the efficacy of this additive to protect a lubricated contact. The standard SRV testing includes conditions that could be inappropriate for evaluating many anti -wear additives including lead naphthenate. This includes the sliding speed which would create a thicker film preventing the additive function from being studied. The contact stress is also much higher in standard ASTM tests which will influence how the tribofilms are created and in the case of lead naphthenate, the

Document

AMS_2020.pdf

312 higher cont act stresses will create additional wear which will react and consume available lead naphthenate making it unavailable to create a protective tribofilm. From this research, the following conclusions were found. SRV Oscillatory sliding experiments : - At 20°C boundary film testing in a pure oscillatory sliding mode, it was shown that samples with lead naphthenate used as an anti -wear additive outperformed neat MAC samples in all experiments. Under vacuum conditions, the samples with lead naphthenate offered twice the wear protection with 440C performing the best under vacuum conditions. - At 75°C in pure oscillatory sliding testing, all experiments performed better on 52100 steel than 440C which vacuum tests showing significant wear reduction . - At 75°C in atmospheric conditions , the neat MAC oil outperformed the samples with lead on 52100 steel with additional concentration of lead increasing the wear rate. Under vacuum, the addition of lead reduced the wear rate. - At both 20°C and 75°C in boundary film testing in a pure oscillatory sliding mode on 440C, samples with lead offered up to twice the wear protection over the neat samples in a vacuum environment. - Lead Naphthenate had better anti -wear performance at lower temperatures and increasing the amount of lead directly reduced the wear rate, apart from 52100 at 75°C and atmospheric conditions . At higher temperatures the wear rate was almost double for the highest lead loading across all tests. MTM mixed rolling/sliding experiments : - At 50°C in the mixed lubrication counter -rotation wear test, lead naphthenate offered three times the wear protection when in a 3% concentration and eleven times the protection for 5% lead compared to the neat MAC fluid. The comparative anti -wear additiv es of TPP produced 20% greater wear and TCP had 3.5 times the wear protection compared to the neat MAC fluid. - At 50°C in the boundary lubrication counter -rotation wear test, lead naphthenate offered three times the wear protection when in a 3% concentration and nine times the protection for 5% lead compared to the neat MAC fluid. The comparative anti -wear additives of TPP reduced wear by five times and TCP had six times the wear protection compared to the neat MAC fluid. - At 150°C in the mixed lubrication counter -rotation wear test, lead naphthenate offered five times the wear protection when in a 3% concentration and twenty times the protection for 5% lead compared to the neat MAC fluid. The comparative anti -wear additives of TPP reduced wear by six times and TCP had two times the wear protection compared to the neat MAC fluid. - At 150°C in the boundary lubrication counter -rotation wear test, lead naphthenate offered three and a half times the wear protection when in a 3% concentration and twelve times the protection for 5% lead compared to the neat MAC fluid. The comparative anti -wear additives of TPP reduced wear by two times and TCP had three times the wear protection compared to the neat MAC fluid. - There appears to be is a transition point between 3% and 5% lead naphthenate where the available lead can react with both the surface and worn metal to create a strong lead anti -wear tribofilm. This tribofilm that is created is between 2- 4µm thick and while it will cause an overall increase in friction at the surface, the wear of the contact is greatly reduced. - Lead Naphthenate had better anti -wear performance at high temperatures and increasing the amou nt of lead directly reduced the wear rate. SOT and Vacuum Bearing Tests: - In both rolling element tests (SOT thrust bearing simulation and ACBB testing), the neat MAC fluid or grease outperformed all samples with anti -wear additives. - The TPP additive performed the best of addatives tested in rolling element tests. - The lead additives performed very poorly in the vacuum bearing tests with greater than forty times the wear of the neat MAC fluid for the oils and twelve to fifteen times the wear in the greases. In the SOT testing, lead additives reduced the relative lifetime by 50- 75%.

Document

AMS_2020.pdf

313 Lessons Learned: - The additive function of lead naphthenate is a combination of physical absorption through rubbing/ pressure and chemisorption. Bearing or t est run-in has been shown to be critical to the generation of proper tribofilms and extending the life of the lubricated contact . - Depending on the mechanics in the contact (sliding versus rolling), the effect of temperature had a significant effect. This appears to indicate that higher concentrations of lead would be required for more severe applications involving pure sliding and/or high temperatures as the lead is consumed faster through reaction with surface layer steel oxides and sublayers. - The formation and durability of the lead tribofilm is dependent on the environment with higher performance coming under vacuum conditions. It is also believed the lack of oxygen promotes this life with reduced oxide formation on the steel and degradation of the lead naphthenate. Acknowledgements I would like to thank the team at Nye Lubr icants specifically Melissa Larochelle, Paul Moses , Richie Raithel, and Mason Wood for the tribology testing and profilometry . References [1] Jason Galary. Mechanochemically Created Tribofilms and the Mechanics of Rolling Wear. Ph.D. thesis, University of Massachusetts at Dartmouth, 2019. [2] H Blok. Gear wear as related to viscosity of oil. JT Burwell, Jr., American Society of Metals, Ohio, pages 199- 277, 1950. [3] HC Mougey and JO Almen. Extreme pressure lubricants. Proc. API, Refi ning Div, pages 76- 94, 1931. [4] FF Musgrave. The development and lubrication of the automotive hypoid gear. J. Inst. Pet, 32:3244, 1946. [5] Jason T Galary. Synthetic gear lubricants go green. Gear Solutions, 8(84):24 31, 2010. [6] Jason Galary. Study of wear properties of environmentally friendly lubricants for gearing applications as a function of film thickness transition. In Environmentally Considerate Lubricants. ASTM International, 2014. [7] Jason T Galary. Efficacy of Lead Naphthenate for Wear Protection in Mixed Lubrication Regime. Proceedings of the 44th Aerospace Mechanisms Symposium, 2018. [8] Bernard J Hamrock and Duncan Dowson. Ball bearing lubrication: the elastohydrodynamics of elliptical contacts. 1981. [9] St. Pierre , Nicole. “Us e of 3D Optical Profilometry to Differentiate Between Additive Chemistry” STLE Annual Meeting, 2017 [10] Didziulis, Stephen V., and Paul D. Fleischauer. "Chemistry of the extreme- pressure lubricant additive lead naphthenate on steel surfaces." Langmuir 7.12 (1991): 2981- 2990. [11] Carre, D. J., P. A. Bertrand, and J. R. Lince. Lead naphthenate additive tribochemistry in hydrocarbon oils . No. TR -2002 (8565) -2. AEROSPACE CORP EL SEGUNDO CA LAB OPERATIONS, 2001. [12] Peterangelo, Stephen C., et al. "Improved additives for multiply alkylated cyclopentane‐ based lubricants." Lubrication Science 25.1 (2008): 31- 41. [13] Marc Ingram, Clive Hamer, and Hugh Spikes. A new scuffing test using contra- rotation. Wear, 328:229 240, 2015. [14] Mangolini , Filippo, Antonella Rossi, and Nicholas D. Spencer. "Influence of metallic and oxidized iron/steel on the reactivity of triphenyl phosphorothionate in oil solution." Tribology international 44.6 (2011): 670- 683.

Document

AMS_2020.pdf

315 Bearing Anomaly for the Sentinel 6 Supplemental Calibration System Gale Paulsen*, Dylan Van Dyne* , Fredrik Rehnmark *, Phil Chu*, and Ted Iskenderian** Abstract The Supplemental Calibration System (SCS) was designed to rotate an elliptical reflector for the Sentinel 6 Advanced Microwave Radiometer (AMR). By rotating the reflector, the system can perform calibrations by comparing the reflection from space to the reflection from an onboard calibration target. During the qualification testing of the SCS, a flaw was discovered that resulted in apparent brinelling of bearings that support the rotating reflector. An extensive investigation ensued to determine the root cause of the problem. Though the most likely root causes of the presumed brinelling ended up being the result of undersized bearing clamp rings and thermal mismatch, there were requirements that drove these design choices. This paper describes the design of the bearing assembly, requirements that drove the design, and documents a challenging investigation filled with ambiguous test results. Introduction and Assembly Overview Sentinel 6 is a spacecraft designed to continue the long- term continuity of satellite altimetry for sea surface height. This mission is a multi -agency collaboration between ESA, EU, EUMETSAT, NASA -JPL, NOAA , and CNES [1]. One of the upgrades to the Sentinel 6 microwave radiometer is a calibration system capable of maintaining measurement accuracy to within a few centi meters over the life of the mission [2]. To achieve this accuracy, the microwave radiometer is equipped wi th a Supplemental C alibration System (SCS) . This system helps correct signal drift by directing the radiometer signal to two different calibration targets of known temperatures. It is the design and testing of this system, which is the focus of this paper. The SCS is mounted to the Reflector Support Assembly (RSA). It is shown as a blue assembly in Figure 1 . Elements that comprise the SCS include a rotati ng Secondary Mirror (Figure 2); a stationary Space View Mirror to provide a “cold” calibration target; a Warm Calibration Target (WCT) to provide a “warm” calibration target; a structure to support the SCS components and t o interface to the RSA; a Standard Dual Drive Actuator (SDDA) (Figure 3) to rotate the Secondary Mirror; a Launch Lock Mechanism to support the mirror during exposure to the launch environment; and a Feed Horn (not shown) to direct the microwaves to a wave guide assembly. In the standard operating mode, the system is positioned in its “science position” where it is collecting information on sea surface he ight. In this case, the Secondary Mirror is directed at the larger reflector on the RSA. To maintain signal accuracy, the system is required to perform a calibration approximately once every five * Honeybee Robotics Spacecraft Mechanisms Corporation, Ensign Bickford Industries, Altadena , CA ** Jet Propulsion Laboratory, California Institute of Technology , Pasadena, CA Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 2020 Figure 1. SCS location on the Sentinel 6 Spacecraft

Document

AMS_2020.pdf

316 days. When the calibration sequence is executed, the Secondary Mirror simply rotates to a hard stop at the WCT position and then returns to a hard stop at its nominal science position. During this process, the Secondary Mirror sweeps past the Space View Mirror, thus collecting the information required to correct for drift in the microwave radiometer signal. The stroke required to rotate fr om the science position to the WCT position (hard stop to hard stop) is approximately 120° . Driving Requirements The primary requirements that drove the design of the SCS bearing and housing materials and arrangement are shown i n Table 1. Preliminary Design Review (PDR ) and Critical Design Review ( CDR) values are shown to highlight changes in requirement values as the development progressed. Each of these requirements had impa cts on the system design for different reasons. Program and design choices made to meet these requirements ultimately resulted in the production of this paper. Table 1. Driving requirements for the SCS bearing assembly Requirement PDR Value CDR Value Implication Schedule February 2017 July 2017 Relatively short development time Mass < 12.4 kg < 15.4 kg All Aluminum structure Mass of High Melting Point Materials (Titanium) Not Fully Defined < 1200 g Aluminum structure preferred Qual Temperature Range (Op and NonOp the same) -45°C to +55°C -45°C to +70°C Routine engineering design Mirror Positioning over Temperature Range <0.2/0.2/0.2 mm in each axis and 0.05/0.05/0.08 degrees including all launch and in- flight effects. <0.2/0.2/0.2 mm in each axis and 0.05/0.05/0.08 degrees including all launch and in- flight effects . Press fits required on both inner and outer races First Frequency Mode > 100 Hz > 100 Hz Mass Sensitivity. More mass on mechanism requires more mass on structure. In the case of mass, the PDR requirement of <12.4 kg was late to flow down which subsequently resulted in a system well over mass allocation. The resultant mass at PDR was 14.2 kg Current Best Estimate (CBE) with 16.3 kg i ncluding Mass Growth Allowance (MGA). At PDR, it was apparent that the 12.4 kg requirement was going to be a significant challenge to meet going into CDR. A compromise was achieved by increasing the mass allocation to 15.4 kg, and the design was scrubbed t o enter the CDR with a CBE mass of 12.8 Figure 2. SCS Components

Document

AMS_2020.pdf

317 kg. The estimated MGA brought the total to 13.7 kg. The estimated MGA was validated when the assembl y’s mass ended up at approximately 13.4 kg. This final mass even included a deviation after CDR that added more mass to the interface plate between the SCS and the RSA to increase the stiffness of the interface. A limit on the use of high melting point materials limited the mass of titanium that could be used in the design. For thermal reasons, a titanium bracket was used as an installation bracket for the microwave feedhorn. The mass of this bracket was only around 150 g. Other titanium parts throughout the design added roughly another 150 g for a total titanium mass of 300 g. In hindsight, there was room in the mas s allocation to add titanium mass to other areas such as bearing shafts and housing s. However, adding more cantilevered mass to the design reduced the natural frequency of the system. Also, preliminary test results , described later, provided confidence in the design choices made through CDR . At PDR, the q ualification temperature range defined was relatively benign; ranging from - 45°C to +55°C. Prior to CDR, this range was extended with the upper limit moving to +70°C. Nevertheless , the thermal environment was not considered a huge obstacle for the mechanism. Alignment of the mirror over the temperature range ultimately played into tolerancing and fits on the SCS bearing assembly. The temperature range over which the positioning requirement applied was reduced from qualification to Flight Acceptance (FA) temperatures. This change was a mitigation approach to reduce the preload stresses in the bearing by reducing the magnitude of interference fits across the temperature range. Meeting the first mode requirement was by far the most challenging part of this design effort. Margin of 20 Hz above the first mode frequency was design ed into the system to ensure the SCS could meet the 100 Hz first mode requirement once integrated into the larger system. The mass of mechanisms and mirrors made 12 0+ Hz a difficult target . This is because the cantilevered nature of this system drove the center of gravity away from the interface ( Figure 3). This led to a constant battle between function, mass, and stiffness. Additional design challenges came from a relatively high random vibration requirement at the SCS interface which started at 16.8 grms , but later reduced to 14.1 grms shortly before CDR . Risk Reduction Effort To proactively reduce risk , Brassboard versions of the thin section bearing assemblies were designed, procured, assembled, and tested. The primary objective of this effort was to reduce uncertainty in the drag torque of the bearing assembly across the operating temperature range. This was important because the control approach for this mission was to perform open loop control over the rotation of the Secondary Mirror using a fixed voltage input into the motor , and s cience was sensitive to the rotational velocity of the Secondary Mirror . Because of this, a require ment on the velocity was written to specify an operational band that ranged from 1.5 deg/s to 2.9 deg/s. Predictions on velocity prior to PDR showed little margin compared to the velocity requirements over the operating temperature range of the system ( Figure 5). With so little margin against velocity , predictions on friction were important to validate. The Brassboard assembly consisted of the output bearing assembly, input bearing assembly, torque transducer, motor and structure (Figure 4). Config uration of the output bearing assembly (designed to support the mass of the Secondary Mirror) was a back to back duplex pair of angular contact bearings with a spring preloaded outrigger radial bearing to help support moment loads. A back to back duplex pair of angular contact bearings were also selected for the input bearing assembly. In this case, these bearings were separated by match ground spacers . Requirements on the input bearing assembly were much more benign because this assembly did not have to support the mass or position requirements of the cantilevered Figure 3. Center of Gravity of SCS

Document

AMS_2020.pdf

318 reflector. In the final SCS configuration, the input assembly simply supported a pinion gear that was mated with a gear on the output assembly. Gears were not included in the B rassboard assemblies. Figure 4. Brassboard test assembly. The test plan for the B rassboard was designed to measure the sensitivity of the design to the thermal environment, including thermal gradients across the bearing races . Multiple tests were performed at room temperature and qualification hot and cold temperatures to measure bearing drag ( Figure 6). Additional data were collected during one of the thermal ramps where a relatively large (~12°C) gradient was enforced across the bearing races . For these analyses, ORBIS bearing analysis software [3] was used to predict bearing stresses and drag over the temperature range. Overall, results were very promising as predictions were within 0.1 Nm for most cases. This torque corresponds to less than 0.1 mNm at the motor which puts the value wit hin the noise of the system. Viscosity for the Braycote lubricant was estimated using supplier data and the model defined by ASTM- D341 to interpolate over temperature. Figure 5. Predicted velocity as a function of temperature for two different gearbox efficiencies Development and testing of the Brassboard provided a lot of value going into CDR. Test results corroborated bearing drag predictions to within reason and the bearing assembly was proven to perform within s pecification over the temperature range with no anomalous behavior. F rom an assembly standpoint,

Document

AMS_2020.pdf

319 there was value in validating and practicing cleaning, lubrication, and assembly procedures. It really provided a lot of confidence in the performance of the B rassboard design. Figure 6. Bearing drag as measured and as predicted over as -built and MMC conditions for the Input and Output Bearing assemblies Instrumenting and Testing the SCS EM in TVAC Following CDR, a configuration change was made to the output bearing assembly. There was an incompatibility identified in the Brassboard design between the fit required to allow the radial bearing to float and the mirror pointing requirement. Analysis of the Brassboard configuration pointing requirements was not explicit ly validated with the Brassboard build. There was no physical space for classic solutions to this dilemma, such as mounting the radial bearing in a diaphragm flexure. A floating radial bearing led to a concern that moment loading of the duplex angular contact pair would be higher than expected during the launch environment. With this, the team (Honeybee and JPL) agreed to dispense the radial bearing and change to a more widely -spaced duplex pair, consistent w ith the input bearing assembly ’s configuration (Figure 8). The change enabled a more predictable fit through the thermal environment, and therefore a more predictable constraint on the position of the secondary mirror that the assembly supports. It also enabled less interference fit on the angular contact bearings, because the clearance fit budget was no longer being consumed by the radial bearing. The reduction in interf erence fit subsequently reduced predicted stresses on the bearing pair while still meeting pointing requirements, as later verified by test. The SDDA is capable of so much torque, offering abundant torque margin, that no observable increase in SDDA motor current could be measured when operating the SCS spin mechanism. This led to a configuration change between vibration testing on the EM and Thermal Vacuum testing ( TVAC ). To drive the Secondary Mirror, the motor and torque transducer combination used to measure drag in the bearing Brassboard were reconfigured to fit within the SCS assembly. This change produced valuable , high - resolution drag torque data. Actuation of limit switches (used to confirm end of travel and positions around

Document

AMS_2020.pdf

320 the center of travel) was easily observed with the torque transducer. Operation of the SCS through the torque transducer ended up being key to identifying anomalous behavior in the mechanism. It was also an important tool in the subsequent anomaly investigation. Torque transducers were also used upon completion of each bearing assembly to perform a “run- in” test of the bearings. The purpose was to have the bearing assemblies operating at a “steady -state” torque, for a given temperature, over the life of the mission. These baseline tests provided a means for evaluating performance and identifying anomalies in the assemblies throughout their qualification cycle. Prior to running the first thermal cycle, a baseline set of data was collected at +20 °C (Figure 7). The next step was to be actuated at +70°C. However, there was an interrupt in one of the feedback channels used to control the Ground Support Equipment (GSE) motor. When the SCS returned back to room temperature, repairs were made to the EGSE and another set of data was co llected. Though it wasn’t identified immediately, there was a ripple on a 12-degree period that became apparent at +20 °C . This ripple, in hindsi ght existed following vibration testing, but was amplified after the first ½ TVAC cycle to +70°C. During the subsequent test at - 45°C, this ripple was impossible to miss. Figure 7. EM unit torque profile prior to TVAC (blue), at room temperature after ½ cycle to +70°C (green), and at -45°C (yellow) EM and Flight Model (FM) Design and Analysis To discover the root cause of the anomaly, the test program came to a halt. These test results were far from expected. The Brassboard assembly had shown that there “should” have been no reason for concern going into TVAC. It validated both performance and analyses. So, what happened? Something was missed, but what?

Document

AMS_2020.pdf

321 Figure 8. Cross -section view showing bearing mount geometry Bearing analysis was performed using Orbis [1] . For all load cases, predicted contact stresses were lower than the applicable operating and non- operating limits per GSFC -STD-7000A for 440C steel bearings (335 ksi (2310 MPa) and 400 ksi (2760 MPa) , respectively , for non- quiet applications ). Alignment across the temperature range was achieved by controlling the maximum interference between the bearing races and mounting surfaces (shaft and housing). Unloaded balls were predicted for the launch load cases at the 3sigma level but none were predicted at the 1- sigma level. Areas of concern were identified at instrument CDR and closed out with additional analysis. There is still uncertainty related to these aspects of the design, which are not easily modeled. None have been ruled out as contributing factors to the torque spikes observed during thermal testing. Neither has any been singled out as the single root cause. The driving load case for non- operational contact stress is launch vibration. Static equivalent loads were generated for design purposes for 3 cases: Random X, Random Y and Random Z (all 3 sigma). A single enveloping load case was constructed using the maximum components and a factor of safety of 1.3 was applied to the l oad ( Table 2). A 1-σ load case was also constructed. Table 2. Enveloping static equivalent load case for launch vibration (3 σ and 1 σ ), including 1.3 FoS Case Fx (lbf) Fy (lbf) Fz (lbf) Fyy (in -lbf) Fzz (in -lbf) 3 sigma 397 539 390 1755 1246 1 sigma 132 180 130 585 415 Bearing analys es presented at CDR predicted healthy load margins (~ 0.9) against contact stress allowables when subjected to the 3- sigma enveloping load case. This is consistent with static load ratings from the manufacturer ( Table 3). Unloaded balls were indicated but no truncation. No unloading was indicated for the 1 σ enveloping load case. Contact stress maps for the 3- σand 1- σ load cases are presented in Fig. 9. Table 3. Bearing Load Rat ings from Manufacturer Silverthin Part # Physical Specifications Static Load Ratings Dynamic Load Ratings Bore Dia. Outside Dia. Pitch Dia. Width Ball Dia. No. Balls Contact Angle Radial Axial Radial Axial in in in in in deg lbf lbf lbf lbf SSB040BU5Z 4.0 4.625 4.313 0.313 0.156 58 30 3310 5890 1400 2100

Document

AMS_2020.pdf

322 Figure 9. Ball bearing contact stress maps for enveloping random vibration load case (3 sigma and 1 sigma) This condition was deemed acceptable in light of guidance per NASA -STD-5017A [4] which states: Bearings should be preloaded with a load calculated to withstand the operational environments with no unloaded balls, known as “gapping.” Gapping under operational conditions is undesirable but may be tolerable in certain cases. However, increased component testing that verifies performance in this condition becomes necessary because it is difficult to pr edict the effects of gapping analytically. Testing should demonstrate lubricant lifetime, bearing component lifetime, specified functional performance and shaft stiffness. Under non- operational environments, it may be permissible to have some balls unloaded . Because the analysis failed to predict the torque anomaly encountered in test, a sensitivity analysis was performed to investigate the effect of varying bearing dimensional parameters and other analysis inputs. The list of cases considered and the resul ting effect on maximum mean Hertzian contact stress is presented in Table 4. The highest stress (283 kpsi/1950 MPa) was predicted for Case 11a in which bearing internal preload is lost. This could occur if the flexible clamp holding the bearings together is defeated by excessive axial loads. The clamp was designed taking into account only the axial component of the launch loads. Accordingly, the pre- set bearing internal preload was selected to be 306 lb (1.36 kN) [ 5], corresponding to the axial load Fx without the 1.3 FoS. The clamp was sized to produce 3X this force (918 lb/4.08 kN ) to press the inner races of the bearings together [6]. A higher clamp force would have required increasing the size of the clamp and the fasteners securing it, which would have exceeded the volumetric constraints on the mechanism design. The bearing spacers were precision ground to ensure the same internal preload in the assembly that the manufacturer ground into the bearings. The clamp design failed to consider the axial resultant of the moment loading on the bearing assembly during launch, which is significant. When this is included for Case 1, the clamp must resist a total axial force of 1056 lb (4.70 kN) (neglecting friction at the mount interfaces, which would tend to reduce the required force). For Case 2, the total axial force to be resisted is 532 lb (2.37 kN) . This force could cause the bearings to separate and result in additional unloaded balls and more concentrated loads in the bearing. Comparing Case 1 and Case 11a, this is what the analysis predicts. However, a load margin of ~0.6 is still predicted against the non- operating contact stress allowable (335 ksi /2310 MPa) for 440C bearing steel in high precision, low torque ripple applications found in NASA -STD-5017A Table 2. Under these conditions, it is difficult to predict exactly what would happen in test using conventional analysis tools. It is al so difficult to diagnose from the test data whether the bearing races moved during vibration or 0.010.020.030.040.050.060.012 345 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 262728 293031 3233343536373839404142434445464748495051525354555657 58Bearing Loads per Ball (lbf), Row1 & Row2, 3 Sigma Case Thrust Load (X), R1 Radial Load, R1 Thrust Load (X), R2 Radial Load, R2 0.05.010.015.020.025.012 345 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 262728 293031 3233343536373839404142434445464748495051525354555657 58Bearing Loads per Ball (lbf), Row1 & Row2, 1 Sigma Case Thrust Load (X), R1 Radial Load, R1 Thrust Load (X), R2 Radial Load, R2

Document

AMS_2020.pdf

323 thermal cycling or both. Damage caused during vibration could conceivably appear and disappear as the bearing is thermally cycled and rotated since the balls could track differently under different test conditions. Table 4. Random vibration sensitivity analysis results Temperature The driving requirement for selecting bearing fits was the alignment accuracy for the Secondary Reflector Mirror shown in Table 5. The cantilevered design amplifies the misalignment due to gapping at the bearing. The minimum press fit for the bearing assembly (largest gap at temperature) and resulting misalignment is shown. Table 5. SCS Secondary Reflector Mirror orientation harmonic error requirement and analytical predictions presented at CDR The effect of temperature on the bearing fits was studied extensively during the anomaly investigation and it was found that maximum contact stress reported at CDR was underestimated because the analysis did not take into account the thicker mount cross -section at one of the bearings and the effect of a steel gear mounted to the hub (the hub is the outer race housing) . Orbis can account for varying mount cross -sections but cannot model a bearing mount consisting of two materials with different properties (i.e. aluminum housing on bearing row 1 and a steel housing on bearing row 2) . Pressing the bearings into a steel hub would result in higher contact stresses than pressing into an aluminum hub of the same dimensions. The results of re- running the analysis w ith these corrections are shown in Figure 10. Due to the CTE mismatch between the shaft and hub, the trend of increasing stress versus temperature is reversed. With an aluminum shaft and hub, the hub grows away from the bearing and the s haft grows toward the bearing as the assembly is heated. The steel gear, however, restrains the hub radially and the bearing is, therefore, squeezed at elevated temperature. The highest contact stress is 283 kpsi (1950 MPa) , predicted for a steel hub that is 20 degrees colder than an aluminum shaft at 70 °C. This stress is significantly higher than

Document

AMS_2020.pdf

324 temperature- induced operational stress reported at CDR but comparable to the random vibration case 11a in Table 4 with zero preload. Figure 10. Effect of temperature on max mean Hertzian contact stress in the bearing assembly (bearing installed at 20C) The alignment analysis was also corrected for the effects of the thicker sections and steel gear. At cold temperature, the misalignment would be increased because the steel gear would restrain the hub from compressing the outer race. The resulting misalignment predictions , even at a worst case, are shown in Table 6. The requirement is satisfied in both cases. Table 6. Corrected mirror alignment analysis. (Case 19b Thick Section/Steel Hub) In consideration of these results, a lesson learned is that for a mechanism using thin section bearings with a combination of tight accuracy, significant loads and temperature variability, design constraints should be loosened to allow mounting of the bear ing in a material with matching CTE . Moreover, when the bearing is designed and analyzed as a rigidly preloaded system, it should truly be rigidly clamped if you are to truly trust the analysis . Otherwise, analysis results may overlook effects which are difficult or impossible to model accurately. The modification of the bearing configuration between the Brassboard and the EM/FM1/FM2 is not believed to have been detrimental to the design. In fac t, it was expected to be beneficial because the change reduced predicted bearing stresses. Had the B rassboard experienced the vibration environment, it too may have experienced similar behaviors. Root Cause Investigation The process of the bearing anomaly investigation was informed by a fishbone diagram ( Figure 12), which lays out all reasonable causes for the drag torque spikes in the SCS bearing assembly . The diagram is color -coded to indicate the status of each possible cause. Green indicates that the cause has been exonerated, yellow indicates that the cause cannot be exonerated until disassembly and inspection of the bearing assembly, and red indicates that the cause is implicated in causing the anomaly . The top branch of the fishbone diagram concerns the bearings themselves . Manufacturing defects, poor design, inadequate analysis, or over -testing could all cause the anomalous torque spikes . The hardness of 150170190210230250270290 -50 -30 -10 10 30 50 70Max Mean Bearing Contact Stress (ksi) Shaft Temperature (C)Effect of Temperature on Max. Mean Hertzian Contact Stress in Bearing CDR Thick Section/Al Hub Thick Section/Al Hub 20 deg DeltaT Thick Section/Steel Hub Thick Section/Steel Hub 20 deg DeltaT

Document

AMS_2020.pdf

325 the bearings was exonerated by measuring the hardness of a set of bearings from the same lot as the bearings in the EM unit . Tests resulted in an average hardness of 57.6 on the Rockwell Hardness Scale C (HRC), which was considered in- family with the specified requirement of 58- 60 for the 440C stainless steel bearings . Bearing analysis via Orbis 3.0 indicated that an angular misalignment of 0.00028 in (7 µm) could introduce localized Maximum Mean Hertzian Contact Stress (MMHCS) greater than 335 ksi (2310 MPa) . Both a coordinate measuring machine (CMM) and computed tomography (CT) (Figure 11) scans were used to evaluate this condition. CMM measurements showed a 0.0004- in (10-µm) difference across the bearing diameter while computerized tomography ( CT) scans did not show any signs of misalignment. However, the CT scans only had a voxel resolution of 0.002 in (51 µm) , which is not fine enough to detect the potential misalignment. The CMM measurements could only be made on the SCS secondary mirror attached to the output shaft. The ali gnment was only considered conditionally exonerated until the bearings could be inspected after a tear -down procedure on the EM unit. Quality issues with the bearing could include (but are not limited to) poor workmanship, damaged components, or out-of-specification material . A review was conducted of all quality systems involved with the procurement, assembly, and testing of the bearings . Contingent on future disassembly and inspection, the bearing quality was believed to be adequate. Poorly designed clamp rings or uncontrolled tolerance stack- ups between inner and outer bearing ring spacers in the output bearing assembly could result in Brinelling behavior . Manufacturing errors on bearing ring spacers could reduce preload and result in gapping during test events which could also damage the races . Either result could create the observed torque spikes . All inspection points and assembly measurements were review ed to confirm that the as -built tolerance stack -up was nominal . Shimming and grinding procedures were confirmed to produce the desired preload, which was measured indirectly (by measuring the gap between the clamp and shaft/hub) during assembly of the bear ings. Review of the clamp design analysis with a high- fidelity Finite Element Analysis (FEA) model confirmed good correlation with the load vs. deflection curves seen in in the as -built assembly. Bearing set running torque was measured during assembly and aligns closely with values predicted in Orbis . Axial acceleration endured by the bearing set during Protoflight vibration testing was measured to be 30% less than predicted, indicating additional overhead on bearing design safety factors . Status of manufac turing errors as the root cause is labeled as Conditionally Exonerated until the bearing set c an be disassembled, allowing for closer inspection and more precise measurement (disassembly of the EM bearing assembly has not occurred at this time) . Over -testing of the bearings could of course impart greater loads on the bearings than their design intended. A detailed review of all test parameters and instrumentation was conducted in relation to the Environmental Requirements Document (ERD). The review concluded that all parameters adhered correctly to the ERD and all measurements were taken with calibrated instruments. This cause was exonerated. The analysis performed to generate the bearing design was conditionally implicated in causing the bearing torque spik es. Analysis was performed with three different suites of specialized bearing software (Bearings Figure 11. CT scan of output bearing assembly

Document

AMS_2020.pdf

326 14, Orbis 3.0, and MESYS Version 07/2019) at all as -built and tested conditions, and agreed on the maximum mean Hertzian contact stress at assembly conditions within a reasonable range of approximately 20 ksi (140 MPa) . None of the tools predicted the anomalous behavior observed, indicating either the bearings cannot be properly modeled in the software, or their as -built configuration does not match the analyzed configuration. Therefore, the analys es performed must be called into question. Figure 12: Fishbone diagram All our nondestructive evaluation methodologies are all uniquely limited in their capacity to detect possible defects, so bearing defects and lubrication were considered conditionally exonerated until disassembly of the bearing set could be performed. The bearing set design was conditional ly implicated as a cause for the torque spikes. This set design can be considered unique or unconventional due to the materials and fits used. Such uniqueness emerged from the complicated dimensional, environmental, and structural requirements imposed on t he SCS. Until the bearing set can be properly simulated with the available analytical tools or inspected during a tear -down procedure, its design must be identified as a leading cause of the anomalous behavior. Poor assembly technique of the bearings could contribute to the type of anomalous behavior seen in this bearing set. However, all three SCS units exhibited the anomalous torque spike behavior at exactly the same position (where mirror was constrained during vibration testing and temperature ramps in TVAC). Although this paper focuses on the EM bearing set, both the Flight Model 1 and Flight Model 2 exhibited a similar behavior. The units only differ in magnitude of the torque spikes, where the EM had by far the highest measured torque spikes. The primary difference in testing between EM and FM1 is how the thermal environment was controlled in TVAC. Thermal gradients across the entire SCS structure were more closely controlled for the FM assemblies. Contamination of the bearing sets was conditionally ex onerated as a cause for the anomalous behavior. Typically, any damage to a bearing set as a result of Foreign Object Debris (FOD) would manifest as a behavior isolated to one build of a bearing set. In this case, however, all three bearing sets that were built and tested exhibited very similar behavior. This cause is conditionally exonerated until a careful disassembly of one of the bearing sets could reveal some type of FOD generated by the assembly process.

Document

AMS_2020.pdf

327 The environments experienced by the bearing sets appeared to be the catalyst for creating and/or exacerbating the torque spikes . Vibration loads, bulk temperature changes, and temperature gradients across the bearing assembly, in particular, are conditionally implicated in causing the torque spikes. Pretest handling was exonerated because functional tests performed after assembly did not result in detectable torque spike behavior. Table 7: Summary of bearing anomaly tests Test Index Test Name Test Unit Environment Results 1 Pre TVAC Torque Measurement EM Room Average measured torque aligned with predicted, small (~0.02 Nm) periodic torque spikes observed 2 TVAC Qual 1st Cycle Hot EM +70°C, Vacuum Torque spike magnitude increased to ~0.05 Nm from predicted drag torque 3 TVAC Qual 1st Cycle Room EM Room Torque spike magnitude appears to be permanently increased at the same locations 4 TVAC Qual 1st Cycle Cold EM -45°C, Vacuum Torque spike magnitude increased to ~0.30 Nm from predicted drag torque 5 TVAC Qual 2nd Cycle Hot EM +70°C, Vacuum Torque spike magnitude appears to be permanently increased to ~0.30 Nm from predicted 6 TVAC Qual 3rd Cycle Hot, Offset EM +70°C, Vacuum Performing hot TVAC cycle with bearing set rotated 6° from nominal position creates new sp ike location 7 TVAC Qual 3rd Cycle Room, Offset EM Room Torque spike magnitude appears to be permanently created at the new locations 8 TVAC Qual 3rd Cycle Cold, Offset EM -45°C, Vacuum New torque spikes still present, no apparent increase to magnitude (~0.05 Nm) 9 TVAC Qual 3rd Cycle Room, Offset EM Room Torque spikes are permanently created following TVAC cycles, not exacerbated at cold 10 Post Vibe Torque Measurement FM1 Room Small torque spikes observed in data, on the order of 0.02 Nm in magnitude 11 Bearing Run- in Test (200 cycles) EM Room Torque spikes diminished in magnitude, from ~0.20 Nm to ~0.05 Nm, 6° spikes are no longer present 12 Pre TVAC Torque Measurement BB Room Brassboard unit exhibits torque spikes with ~0.04 Nm magnitude 13 TVAC Qual Cycle Hot BB +70°C, Vacuum Torque spike magnitude increased to ~0.10 Nm 14 TVAC Qual Cycle Room BB Room Torque spikes appear permanently increased from previous hot cycle 15 Pre TVAC Torque Aluminum Gear EM Room Torque spikes (~0.06 Nm) still present in bearings with steel gear replaced with aluminum stand -in 16 TVAC Qual Cycle Hot EM +70°C, Vacuum Torque spikes increased to ~0.10 Nm with aluminum gear 17 TVAC Qual Cycle Room EM Room Torque spikes permanently increased, average running torque also increased 18 TVAC Qual Cycle Cold EM -45°C, Vacuum Torque spikes still present, no noticeable change in magnitude 19 TVAC Qual Cycle Room EM Room Torque spikes still present, no noticeable change in magnitude, running torque in -line with pre hot cycle 20 TVAC Qual Cycles Repeat at Hot EM Room & +70°C, Vac Repeated tests with steel gear reinstalled at hot and room temperature shows no significant change to torque spike magnitude The test program includes over 20 discrete thermal vacuum tests meant to collect data on the bearing sets’ anomalous behavior and find a root cause. The types of tests run were informed by various hypotheses developed between Honeybee and JPL during the investigation. Table 7 summarizes the tests performed.

Document

AMS_2020.pdf

328 For the sake of brevity, only the plots from a selection of these tests are shown in accordance with their potential to cause the anomalous behavior. Conclusion s The anomaly in the SCS bearing assembly has been humbling, but interesting. A Brassboard assembly was built and tested to provide validation of the design concept and tools. Deviations to the design were made between the Brassboard and EM/FM designs that were expected to reduce stresses in the bearing assembly. Test results have shown that there was design oversight in the process used for the SCS bearing assembly. The root cause investigation eliminated several potential contributors or reasons for the anomaly. However, the team is still left with questions. • Did the anomaly occur during vibration testing, but only appear during TVAC due to some shift in the position of the ball track? • Is the anomaly a function of temperature or temperature gradients only? If so, why isn’t drag in the system unexpectedly high for the full stroke of the mechanism? Average drag actually remained reasonably consistent with predictions. • Is friction between the bearing and the shaft and housing too large to ensure the intended clamping over the temperature range, given the CTE mismatch? • Was the problem with the clamp stiffnes s? If so, this would indicate that the anomaly occurred during vibration testing as non- rigid clamps would only r educe stresses during thermal cycling. There are currently no plans in place to perform more tests, inspections, disassembly, or otherwise to find the true root of the problem for this assembly. However, there are some good lessons learned, or good reminders of lessons that have been learned previously by others. These are true particularly for thin section bearings which can be more sensitive to design parameters. • Fundamentally sound designs should be the starting point. In this case, use matching CTE materials , “rigid” clamps , and avoid mixed material designs that are difficult to model (e.g. the gear in the SCS assembly was restricting the growth of the housing). • Question requirements if they are dictating solutions that are difficult to analyze and push back if necessary . If the opportunity exists to pursue this investigation further there are more tasks that could be performed to help identify the root cause. These include disassembly of the EM to inspect the raceways; and assembly and test (both vi bration and TVAC) of a few more units with changes to specific variables such as clamp stiffness, housing and shaft friction, and CTE mismatch. Acknowledgments This research was carried out at the Jet Propulsion Laboratory, California Institute of Technolo gy, under a contract with the National Aeronautics and Space Administration (80NM0018D000 4), subcontract 1531184. References 1. ESA 2000 -2019, “Sentinel -6 Mission.” https://sentinel.esa.int/web/sentinel/missions/sentinel -6. Accessed December 23, 2019. 2. ESA 200 0-2019, “Sentinel -6 / Jason-CS.” https://directory.eoportal.org/web/eoportal/satellite- missions/content/ - /article/jason-cs. Accessed December 23, 2019. 3. Halpin, Jacob, D., “Ball Bearing Analysis with the ORBIS Tool,” Proceedings of the 43rd Aerospace Mechanisms Symposium, NASA Ames Research Center, May 4 -6, 2016. 4. NASA Technical Standard. “Design and Development Requirements for Mechanisms.” NASA -STD -5017A w/CHANGE 1. 2016 -05-31. 5. Singer, H., “Space Vehicle Mechanisms: Elements of Successful Design,” ed. Conley, P., John Wiley & Sons, 1998. Ch. 12, p. 305. 6. Videira, et al. “Design, Assembly and Preloading of Ball Bearings for Space Applications – Lessons Learned and Guidelines for Future Success.” Proceedings of ESMATS. Noordwijk, The Netherlands . 25-27 September 2013.

Document

AMS_2020.pdf

329 Parker Solar Probe MAG Boom Design, Analysis and Verification Weilun Cheng*, Calvin Kee* and John Wir zburger* Abstract For the “ hottest and coolest ” mission to the Sun, FIELDS is the science instrument suite on the Parker Solar Probe (PSP) spacecraft that measures the magnetic fields in the solar corona during close approach to the Sun. The FIELDS instrument suite features five electric antennas and three magnetometers . The MAG netometer (MAG) boom accommodates all three magnetometers and one of the electric antennas, the V5 antenna. The unique PSP sun shield – known as Thermal Protection System (TPS) – generates an umbra for the spacecarf t and instruments so that the temperature environment is manageable during close approach to the Sun. The cone- shaped umbra tapers off towards the bottom of the spacecraft. With the tip of the deployed boom 3.5 meters aw ay from the bott om of the spacecraft , the umbra imposed a tighter alignment requirement for the deployed boom than the instrument pointing requirement . This paper describes the design, development and verification of the Mag boom. Included are the dynamic analys es of boom deployment using Ad ams software, and the verification of the final on-orbit boom deployed position using the V5 antenna, solar limb sensors, and spacecraft attitude maneuvers. Introduction Parker Solar Probe is a mission to touch the sun. It is the first- ever mission to get as close as 6.16 million kilometers to the sun. The mi ssion is planned for seven years, including 24 orbits around the sun using seven Venus flybys to gradually reduce the perihelion distance relative to the sun. The spacecraft is about the size of a small car. T he thermal protection system on top of the spac ecraft is made from a 11.43- cmthick carbon- composite structure which can withstand temperatures of up to 1,400 degrees Cel sius and beyond . There are four major investigations for the mission: 1) Fields Experiment ( FIELDS) is designed to measure magnetic fields in the solar corona; 2) Integrated Science Investigation of the Sun (ISOIS) will make observations of energetic electrons, protons, and heavy ions that are accelerated to high energies; 3) Wide -field Imager for Solar Probe (WISPR) will take images of the solar corona and inner heliosphere, solar wind, shocks and other structures as they approach and pass the spacecraft; 4) Solar Wind Electrons Alphas and Protons (SWEAP) Investigation will count the most abundant parti cles in the solar wind and measure their properties such as velocity, density, and temperature. This paper focusses on the magnetometer boom that accommodates four of the FIELDS instruments. MAG Boom Description The Parker Solar Probe Mag boom measures 3.5 meter s long. It has two saloon- door style hinges, two composite boom segments, four instruments, two connect or brackets, and one harness bundle, weighing a total of 9.02 kg. It was restrained to the spacecraft with two launch locks, the +Z loc k and the - Z lock. The two boom segments were made from a layup of M55J/RS3C uni -directional carbon- fiber cyanate ester prepreg. 6AL-4V Titanium end fiitings were bonded to the boom segments with Henkel Hysol EA -9360 adhesive. The four instruments included two Fluxgate Magnetometers (MAGi and MAGo) , one Search Coil Magnetometer (SCM) , and one V5 antenna. They were all populated on the outward segment of the boom assembly to allow maximum distance from the spacecraft while staying within TPS shadow , the umbra . Each instrument came with a harness pigtail that connected to either of the two connector brackets based on where each instrument was located. This a rrangement allow ed last-minute instrument replacement , if * Johns Hopkins University Applied Physics Laboratory , Laurel, MD Weilun.Cheng@jhuapl.edu , Calvin.Kee@jhuapl.edu , John.Wirzburger@jhuapl.edu Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 2020

Document

AMS_2020.pdf

330 there was a need, without disturbing the test verified harness bundle loops and restraint s across both hinge joints. There was a G -10 spacer, coated with BR -127 ESD, placed under the shoulder hinge to thermally isolate the deployed boom from the spacecraft. See Figure 1 for MAG boom description, and Figure 2 for deployed boom and umbra. Figure 1. MAG Boom Assembly Description Figure 2. Deployed Boom and Umbra TiNi FC3 frangibolt s were used as the release actuator s for both +Z lock and –Z lock. Each lock assembly used a Viton pad as a shock damper. There was also a preloaded spring plunger at the +Z lock to provide a push- off force at the separation interface during deployment. To mainta in Mag boom magnetic cleanliness, the plunger was installed on the launch lock housing that mounted on the spacecraft. There was no spring plunger at the –Z lock since the outward boom was preloaded at the –Z lock. The boom spring- back force after frangibolt firing was much higher than the push off force from the plunger. On orbit, after firing the +Z lock and –Z lock in sequence manually , the Mag boom swung away from the spacecraft and deploy ed toward the bottom of the spacecraft . It went through a transient phase of over -travel and under -travel and settled to the final position. The final Mag boom position was controlled by the hinge holding torque. Similar ly design ed hinges were used on multiple previous APL missions with excellent , successful heritage [1]. Figure 3 illustrate s the Mag boom deployment sequence . Figure 3. MAG Boom Deploy ment Sequence

Document

AMS_2020.pdf

331 Ground Test Verification In the early program phase, an Engineering Model (EM) boom was built to demonstrate the design concept, the hinge performance, the harness routing scheme, the boom deployment transient behavior and the deployment settling time. Multiple standalone in-air deployment test s on an air-bearing table sucessfully demonstrated the Mag boom performance. Qualification level environmental tests performed using the EM boom validated the structure and thermal design margins , and also demonstrated compliance with the alignment requirement s. Figure 4 lists the EM boom test sequence, i ncluding hinge test, bond joint strength test, boom segment bend test, boom assembly tests, as well as alignment verifications. Figure 4. EM Boom Test Sequence The F light Model (FM) boom was built under extreme control for maintaining magnetic cleanliness. Only magnetic -clean certified tools we re allowed in the designated Mag boom work area. Acceptance level environment tests validate d the FM boom workmanship. Standalone in-air deployment test ing using flight frangibolt actuators verified the boom deployment sequence and deployed boom alignment . Figure 5 shows the FM Mag boom standalone deployment test. Figure 6 lists the FM boom test flow. Figure 5. Flight Boom Deployment Test

Document

AMS_2020.pdf

332 Figure 6. FM Boom Test Flow After integrating to the spacecraft, the Mag boom was included in some spacecraft level tests , including pop-and-catch followed by partial walk out, mass property, vibration, acoustic, thermal balance, and full walk out tests. Figure 7 lists the Mag boom related spacecraft level test flow. Figure 7. FM Boom Related Spacecraft Test Flow The spacecraft full boom walk out test was performed to verify that the full boom range movement, plus over travel motion, was free of obstr uction from Multi- Layer Insulation ( MLI) blankets , harnesses, tapes, and other surrounding components. With the spacecraft lifted in the air and the boom positioned below the spacecraft, the test introduced a major safety concern. The final test approach was to maneuver the boom using extension rods so operators c ould avoid working under the suspended load. The test procedure was rehea rsed several times using the EM boom and a spacecraft mock up. T he final full walkout test was successfully performed on the spacecraft with no abnormal findings . See Figure 8.

Document

AMS_2020.pdf

333 EM Boom Practice Spacecraft on GSE D olly FM Boom Test Figure 8. Boom Full Walkout Test on Spacecraft Dynamic Analysis of the Mag Boom Deployment Deployment Analysis Background The Mag boom assembly mechanical deployment was demonstrated and verified at various phases, from the engineering model stage to flight model stage. The critical verification was the final , flight -like integration level, flight model stage where all flight components including harness, harness connectors, harness holdowns, thermal blankets and micro -meteoroid shielding were assembled onto the Mag boom, along with mass simulators in place of flight instruments. Test verification of deployment was the confirmation of the two Mag boom arms swinging out and settling into their design positions, which were validated with precision alignment measurements. In the preliminary design phase, a boom- deployment analytical model was built using Adams multi -body dynamics software. The model was not very detailed in complexity, but was sufficient to provide an indication whether or not the dual -arm Mag boom concept would work for the Program. The deployment simulation based on this model provided a level of confidence that the sweep of the arms would not impact the spacecraft, and that the torque at the hinges would be adequate to deploy the Mag boom to its intended position. Another Adams model was built in the later phase, after the final flight -like integration and testing phase, to closely represent the final flight deployment test behavior of the Mag boom. The objectives of generating a test-verified model were to observe the dynamic effect of deployment of the Mag boom on the spacecraft, and to have the capability of running various scenarios of spacecraft maneuvers in orbit and other dynamics related predictions or validations. The Adams model can provide a validation, or at least an indication, of the dynamic measurements received from the spacecraft.

Document

AMS_2020.pdf

334 Description of Adams Model The primary Mag boom assembly geometry was imported into the Adams model. The analytical model consisted of the shoulder hinge, the inner boom, the elbow hinge, the outer boo m, the primary launch lock and snubber, the secondary launch lock, and the four instruments that mount on the outer boom ( Figure 9). Harness and blankets were included as smeared masses on the two boom segments. The flight -like Mag boom that was installed on the deployment, gravity -negated table was weighed so that the Adams model mass could be adjusted to match. All the components in the Adams model were treated as rigid bodies, each having its own mass moment -of-inertia and center -of-mass. Since the Mag boom in the deployment test was oriented with hinge axis being vertical, and the booms were supported on the deployment table with compressed air bushings, it was assumed in the Adams model that gravity was negated. Figure 9. Adams Multi -Body Dynamics Model of Mag boom There were three boundary constraints in the Adams model. The shoulder hinge bracket which tied the Mag boom to the spacecraft was fixed in 6 degrees -of-freedom (DOF). The primary launch- lock bracket on the inner boom, close to the elbow hinge, held the other stowed end of Mag boom in 6 DOF. The secondary launch- lock bracket at the tip of the Mag boom was also fixed in 6 DOF. These launch -lock constraints were timed to be removed in the solver in order to release and deploy the Mag boom. A preload force was applied at the snubber between the booms when the outer boom folded over the inner boom in the stowed configuration. The force was derived from a flexible Nastran static analysis of the Mag boom . This force was a timed reaction between the two booms, and as the outer boom moved apart from the inner boom, the force went to zero. Hard stops in the rotations of the shoulder and elbow hinges were modeled as stiff impact -contacts. To provide the work required to open each hinge, a spring torque was applied at the shaft of the hinge. This torque was not defined as a typical Adams built -in torsion spring with a stiffness and a preload. Because of the nature of the “saloon- door” hinge, which was sprung i n both directions, an equation had to be derived to apply this reversing torque that was based on the angle of the hinge brackets. There was a minimum torque designed into each hinge of approximately 2.26 N ·m, where the minima of the torque curve was situa ted, and that minima corresponded to the angle of the settling position of each boom. A damper was designed into the hinge to reduce the overtravel of the hinges and to absorb some energy before the hinges impacted their hardstops (Figure 10). The componen ts of the damper were included in the multi -body simulation, and to account for the activation of the damper, a rotational spring- damper function was modeled that applied the damping only when the damper brackets had rotated with respect to each other. Friction was also included in the hinge modeling.

Document

AMS_2020.pdf

335 Figure 10. Components of a Mag Boom Hinge as Modeled in Adams Comparison of Deployment Test and Simulation Spring torque, damping, boom kick -off preload and contact stiffness -penetration were adjusted and a simulation that had a close representation of the deployment behavior and duration was progessively developed and established. The main correlation goals for the Adams model were to generate the same deployment shapes, and to complete the deployment clos e to the same duration as the deployment test. Figure 11 shows the Mag boom on the deployment platform and the Adams model prior to actuation of the release mechanism. Figure 12 through Figure 15 capture the major deployment shapes of the Mag boom. Figure 16 shows the final position of the Mag boom after release and settling from its oscillation. The simulation appeared to capture most of the dynamic behavior, in particular the overtravel of each hinge, the trajectory of each arm due to the multiple joints of the boom, and the max swing during oscillation. It was noted, while tuning the parameters, that the time of the shapes in the analysis tended to be quicker in the early part of the deployment, and then ended up taking a longer time to settle. The presumption was that the Adams model did not take into account the air resistance on the Mag boom surface, the friction between the air bearing and the platform, and the additional weight of the air bearings and hoses, all of which would increase the response t ime in the simulation. At the end, where the settling time took longer in the simulation, the lack of air resistance was not impeding the oscillation of the Mag boom. Overall, the Adams simulation was deemed sufficient to describe the dynamic behavior of the Mag boom deployment as it closely followed what was seen in the test. As in the test, the simulation showed that the sweep and reach of the arms would not cause any impact events with the spacecraft. Figure 11. Mag boom Test Configuration and Adams Model Prior to Release

Document

AMS_2020.pdf

336 Figure 12. Release Kickoff, Test at 0.5 sec and Analysis at 0.3 sec Figure 13. Elbow Hinge Overtravel, Test at 3.0 sec and Analysis at 1.2 sec Figure 14. Shoulder Hinge Overtravel, Max Mag boom Swing, Test at 12.5 sec and Analysis at 9.9 sec Figure 15. Mag boom Max Rebound Swing, Test at 21.0 sec and Analysis at 19.0 sec Figure 16. Mag boom Settled, Test at 70 sec and Analysis at 166 sec.

Document

AMS_2020.pdf

337 Prediction of On- orbit Deployment Effect on Spacecraft The Adams model of the Mag boom was integrated with a representative PSP spacecraft , matching the geometry , mass, moments of inertia, and center of mass of the spacecraft bus (Figure 17). The objectives were to observe how the analytical deployment affected the attitude of the spacecraft, and whether the behavior of the Mag boom and the spacecraft would couple together to cause an impact scenario. Figure 17. Mag boom Integrated on Spacecraft Figure 18. Mag Boom Deployment and Effect on Spacecraft Figure 18 illustrates the Mag boom deployment and its effect on the spacecraft. As the Mag boom unfurled, the spacecraft showed very small changes in attitude. When the fully outstretched boom swept toward the

Document

AMS_2020.pdf

338 rear of the spacecraft, there was a larger moment offset from the spacecraft, which caused it to rotate off its original attitude. The largest angle off the spacecraft axis was seen when the Mag boom reached it s maximum overtravel after sweeping past the spacecraft axis . The angle was determined to be 12 deg, and after approximately 12 oscillations, the Mag boom motion damped out, and the spacecraft attitude settled at 7.6 deg ( Figure 19). It was found that the spacecraft angular velocity during the deployment was at a max of 2.6 deg/sec (0.045 rad/sec). Figure 19. Attitude of Spacecraft Off-Axis was Max 12 deg and Settled at 7.6 deg, while Max Angular Rate was 2.6 deg/s On-orbit Deployment and Final Positi on Verification On-orbit deployment of the magnetometer boom began on Flight Day 2, just after 24 hours post launch. Many factors influenced the decision to deploy the boom so quickly after orbit insertion including the blockage of thrusters that needed to be commissioned prior to a trajectory correction maneuver one week after launch, and diminishing data rates as the vehicle traversed farther from Earth. The on- orbit deployment verification consisted of three significant parts : confirmation of the fracturing of the intermediate frangibolt prior to detonating the primary frangibolt for boom motion, confirmation of the primary frangibolt fracturing causing boom motion, and final verification of the position of the magnetometers. Intermediate Frangibolt Detection Frangibolt heating for the Mag boom Hinge Release +Z Tall Side on the primary side was initiated with temperature and timing cut -offs to prevent continuous current draws in the case of a failure. Proper actuation of this frangibolt would not result in boom motion, as the actual deployment consists of rupturing two frangibolts. Even with the nominal plan of heating the frangibolt through both the primary and secondary paths, it was desirable to confirm the breaking of the first frangibolt prior to releasing the second frangibolt. Without motion of the magnetometer boom, it was unlikely that the science magnetometer on the boom could be used for verification of the fracturing of this first frangibolt. Looking at other possible sensors capable of detecting such an event, PSP is equipped with a Northrop Grumman Scalable Space Inertial Reference Unit. This inertial measurement unit consists of four low noise hemispherical resonator gyros as well as four accelerometers, enabling sensing of both rotati onal and translational motion of the spacecraft. Additionally, each set of four sensors is arranged in a tetrahedral, allowing for redundant sensing of events; that is, a glitch on a single sensor can be distinguished from an actual event that would be obs ervable on multiple sensors [ 2].

Document

AMS_2020.pdf

339 As shown in Figure 20, when the temperature of the frangibolt reached its fracture temperature near 70°C at time 75 seconds, an impulsive disturbance was observed across all four gyros. This is indicative of the frangibolt breaking and imparting its energy on the spacecraft. W hile this is not a guarantee of proper actuation, this disturbance level is significantly above the observed quiescent noise floor and the close proximity to the stated fracture temperature is a strong indication. The absence of a disturbance when the redundant current path was exercised reinforces that conclusion or leads one to a double fault scenario, seeing both paths would have not properly activated the frangibolt. Figure 20. +Z Mag Frangibolt Actuation, Gyro Rate Time Histories During ground test ing, some activation signatures were sufficiently small that they were unobservable in the time histories directly. For those cases, the method of using shock spectrums proved to be sufficient. By analyzing how the frequency content in the gyro measurement s changed over time, one could discern the impulsive event as a release of energy across a broad spectrum of frequencies. The on- orbit activation of the +Z Mag frangibolt is shown in Figure 21. The in -flight signature was similar to ground test spectrums. Figure 21. +Z Mag Frangibolt Actuation, Gyro Rate Shock Spectrum With positive confirmation of the +Z Tall Side frangibolt released, it was deemed safe to proceed with the full boom deployment.

Document

AMS_2020.pdf

340 Second Frangibolt Detection The fracturing of the –Z Short S ide frangibolt for the magnetometer boom hinge release occurred in a similar fashion to the +Z Tall Side, without the need to scrutinize the gyro data to determine if the deployment transpired. Like many spacecraft , PSP was mass and power constrained, leading to the selection of small reaction wheels. While the size of the wheels is suitable for nominal operations, the rate of change of the momentum of the deploying boom was sufficient to overwhelm the torque capability of the wheels. Instead of engaging thrusters to counteract the momentum of the deploying boom, since the vehicle was in a benign orbital location from a pointing standpoint, the spacecraft was allowed to be pulled off attitude by the boom deployment and autonomously return to the desired atti tude. This sequence of events is displayed in Figure 22. Figure 22. Attitude Response to Magnetometer Boom Deployment At time 30 seconds, the boom begins to deploy and brings the spacecraft off sun- point, peaking at 11 degrees off its initial attitude. Over the next 60 seconds, the boom motion dampens out, following which a slow recovery slew imparted by the reaction wheels is performed. The complete time from initial boom motion to restoration of attitude was on the order of 180 seconds. Deployed Boom Position Verification The FIELDS Instrument Suite is a set of 5 voltage and 3 magnetic sensors designed to measure DC and fluctuating magnetic and electric fields, plasma wave spectra, the PSP’s floating potential, and solar radio emissions. To achieve this goal, the varying sensors are located across the spacecraft, with four located on the magnetometer boom. These f our consist of two fluxgate magnetometers, a search coil magnetometer and a voltage sensor (V5 Sensor) [ 3]. To survive solar thermal inputs and to function properly, these sensors must be maintained in the shadow of the spacecraft. During commissioning, an activity was planned to verify the coarse positioning of the boom prior to the first solar encounter to ensure the safety of the instruments. It is important to note that while the V5 Sensor is a voltage sensor, it is also photosensitive and can be used to determine if solar input is incident on the detector. To accomplish the verification, vehicle slews moving the Sun out from behind the Thermal Protection System and down each optical axis of the seven sun sensors were performed. Locations of the sun sensors and V5 Sensor are shown in Figure 23. While the slews were designed to illuminate the sun sensors, they

Document

AMS_2020.pdf

341 were not sufficiently large to illuminate a correctly positioned magnetometer boom, nor the bar of the V5 Sensor. Nominal shadow patterns are presented in Figure24. Absence of a Sunlight signature on the V5 Sensor during the slewing profile would indicate the boom was positioned correctly in the umbra, behind the spacecraft. Temperature data was also monitored during the slews. Confirmation of the proper attitude profile was performed using star trackers, gyros, and sun sensors. Following the on- orbit execution of the slews, Gu idance and Control analysts and FIELDS scientists reviewed the data collected and determined that the coarse alignment of the magnetometer boom was nominal, inside the umbra of the Thermal Protection System. This has now been corroborated by three separate solar encounters where no ill effects of solar intrusion on the boom have been noted. Figure 23. Placement of the V5 Sensor and Sun Sensors Figure 24. Shadowed region at the V5 Sensor

Document

AMS_2020.pdf

342 Conclusions The Parker Solar Probe thermal protection system umbra created a unique alignment reqirement for the MAG boom. Multiple G -negated boom standalone deployment tests with both EM boom and flight boom validated boom alignment accuray and repeatbility. The ADAMS model analysis provided a prediction on boom deployment motion, angles, and settling time in space. The on- orbit boom position verfication using V5 instrument, solar limb sensors, and slewing the spacecraft confirmed a successful boom deployment to the nominal position. Lessons Learned • Prior to PSP, all APL saloon -door style hinge deployables were deployed with one hinge line at a time. Deploying two hinge lines at the same time had not been previously performed. With PSP bi - fold Mag boom experience, it is proven that deploying a Mag boom with two hinge lines at the same time is feasible. • Add a position indicator at each hinge joint to provide a positive indication of the final deployed position. • Plan to provide as many blanket grounding points as possible. For the Mag boom, there were 34 groun ding point s available. However, for the two Flexgate magnetometers alone , there were 64 ground lugs to be grounded. Needless to say, there were boom blankets and SCM blanket ground lugs that need to be attached also. It was a big effort to figure out how t o attach all the blanket ground lugs. • Always include safety engineers in the deployment test planning at the beginning. For the full boom walk out test on the spacecraft, there were a lot of safety concerns. Final test strategy was approved by both APL and GSFC safety engineers. • Future efforts to correlate an Adams model to a test setup should include ground- based and ambient characteristics. Such as air resistance, frictions and drags from ground support equipment sources. • Ground test s should include all teams responsible for in-flight monitoring. This will allow the teams first-hand knowledge on how data was collected and how to properly interpret the data that feeds into the analysis . Acknowledgments We would like to acknowledge the support of the APL PSP project management and that of our sponsor, the NASA Goddard Space Flight Center , along with APL Space Exploration Sector line management. We would also like to thank mechanical engineers of NASA Goddard Space Flight Center 5430 br anch who loaned us the GSFC air -bearing tables for deployment testing of the PSP Mag boom. References 1. W. Cheng, et al, Solar Array and High Gain Antenna Deployment Mechanisms of the STEREO Observatory, AIAA 2007- 6096 (2007) 2. Northrop Grumman. (2016). Scalable SIRU Family. Retrieved from Northrop Grumman: https://www.northropgrumman.com/MediaResources/MediaKits/Satellite/Documents/SIRU_Family.pdf 3. S. D. Bale, et al, “The FIELDS Instrument Suite for S olar Probe Plus. Measuring the Coronal Plasma and Magnetic Dield, Plasma Waves and Turbulence, and Radio Signatures of Solar Transients,” Space Sci. Rev. 204, 49– 82 (2016).

Document

AMS_2020.pdf

343 Development of a Low -Shock Payload Fairing Jettison System Boris Halter*, Josef Zeman n*, Simon Wieser*, Beatrice Burkhart*, Mathias Burkhalter* , Alberto Sánchez** and Oliver Kunz* Abstract Today’s robust and reliable separation and jettison systems safely actuate and execute the separation of the payload fairing from the launcher . However, shock- loads are exported from the pyrotechnic separation event . Further more, t he verification of new or adapted fairings is typically carried out by extensive and costly test campaigns due to the installation of full- scale components in vacuum chambers. I n recent years, there is an increasing demand from the launcher primes and sate llite suppliers to improve the payload comfort and to simplify the verification process. Following this need, RUAG Space is developing within the ESA - funded Future Launcher Preparatory Program a low -shock separation and jettison system applicable to currently used and future PLFs. The concept is based on a simple and reliable low -weight system for a controlled separation using hinges and actuators to rotate and jettison the PLF halves. This paper describes the developed concept and the mechanisms for the rotation and jettison. The focus is on the event of separation and jettison, excluding the method for fixation of the fairing during launch and the initial release. The paper shows the advantages of the novel design and explains the verification activities performed to prove the system. Introduction RUAG Space is a leading supplier of products for the space industry in Europe and has as well a growing presence in the United States. RUAG Space develops, designs and builds payload fairings for several launch vehicles of different classes, having decades of experience in reliable and robust technical solutions with 100% mission success. Furthermore, RUAG Space is developing and manufacturing high reliable mechanisms for spacecrafts and launchers , as the herewith in this paper presented separation system. One of the key events in a satellite launch is the separation and jettison of the payload fairing (PLF) from the launcher. In the following a short overview of the state- of-the-art separation system’s main characteristics is given. The central reasons for the use of current fairing separation systems, based on pyrotechnic solutions , include: • High efficiency : the pyrotechnic separation system provides a high ratio of separation force versus system weight. • After PLF separation only a small section of the interface ring remains on the launcher as t he pyro cords are typically encapsulated inside the PLF frame rings that are broken for separation. This influenc es positively the fuel consumption of the mission. • Simple and compact design: pyrotechnic systems are not only compact, but also have a simple design compared to other mechanical systems ; less parts are included and no movable joints or interfaces are needed, increasing the reliability of these systems. • Flight -proven reliability and large heritage of the solutions: the large heritage of the pyrotechnic solutions with thousands of successful separations makes them a safe and reliable system consisting of only few components and fail -safe solutions. * RUAG Space Spacecraft, Zürich, Switzerland ** RUAG Space Launchers, Zürich, Switzerland Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 20 20

Document

AMS_2020.pdf

344 The need f or a new separation system development is linked to the drawbacks of the current system s: • A relatively high shock load during the separation event is transferred to the rest of the launcher structure, including the payload. This necessitates that all the components attached to the PLF, including the rest of the launcher and p ayload, are d imensioned to sustain this load case. • The fairing separation typically takes place at over 80 km of altitude, at the last layers of the atmosphere, with conditions similar to vacuum. For ground testing, however, t he reaction forces from the pyrotechnic separation s ystem are affected by the air resistance, driving the long-term clearances. For this reason, the pyrotechnic separation systems typically need to be tested in vacuum conditions to be fully representative of the separation event. Today there are few locations worldwide where this test can be carried out due to the size of the test items, structures over 20 meters in height. One of them is NASA’s Glenn Research Plum Brook Station in Sandusky, OH. • Complexity of triggering: a drawback is the demand in accuracy to synchronize the horizontal and vertical separation systems in order to control the clearance. Figure 1: RUAG fa iring test at NASA Glenn Res earch Center Plum Brook in San dusky, OH [1] The main focus of th e new development s described in this paper is to overcome the above drawbacks of the pyrotechnic based separation systems used today . The system under development is designed to cover the separation of different PLF classes with low shock, enabling on- ground tes ting while ensuring the reliability and robustness of current pyrotechnic based solutions. The envisioned system consists of a scalable horizontal and vertical separation system (HSS & VSS , not part of this paper ), compatible with a passive jettison system using hinges and actuators that once triggered, rotates and pushes away both PLF halves in a controlled maneuver . The focus of this paper is on the developed mechanisms of the passive jettison system. System Description The within this paper presented mechanisms are developed for the payload fairing’s passive rotation and jettison after the HSS and VSS release. The system consist s of two independent components : • Pneumatic Actuators: to provide the kinetic energy for the PLF rotation. • Preloaded Hinges: define the PLF rotation axis and provide the jettison kick -off energy.

Document

AMS_2020.pdf

345 The functioning principle of the PLF separation system under development includes three phases as visible in Figure 2: 1. Low shock actuated horizontal and vertical separation system (HSS and VSS) , which are not part of this paper. 2. Passive PLF rotation: actuators push the PLFs to a controlled rotation, guided by the hinges. 3. Passive PLF jettison : the hinges release the connection of the PLF to the launcher at a given angle. Additionally, a jettison energy is implemented to the PLFs by pre- loaded springs, in order to ensure clearance. Figure 2: Separ ation System Principle This approach has several advantages: • Signals do not need to be synchronized: Due to the use of passive systems, there is only a single triggering signal to release the HSS and VSS and initiate the fully passive separation. • Modular: The separation and jettison systems are independent , which allows individual customiz ation or even replac ement of each sub-system for a different product without affecting the other subsystem. Additionally , the actuators and hinges can be outlined and tested independent ly. • Scalable: The solution under development and described here is meant for small and medium size fairing and may also be applicable to larger firings as well, keeping the same working princi ple and thus reducing the development effort. • On-ground testable: The non- pyrotechnic separation is slower than pyro- based separations. As a result, the friction with air becomes less dominant, which makes ambient verification possible. • Simpler modelling: The functioning principle allows for a simpler kinematic modelling. • Low shock: The k inetic energy for the jettison is not provided by an instant impulse but by a continuous force from the actuator and the springs inside the hinges. • Clearance control: the entirely guided rotation allows a higher control of the clearance to the launch vehicle. In the following paragraphs the selection and development of the actuators and hinges for the separation system are described separately. Actuators Given the system concept described above, the actuators need to provide the thrust to overcome the a pex of the fairing half shel ls against the launcher’s acceleration loads and provide the kinetic energy to reach separation. Commercial of f the shelf gas springs have been chosen as passive actuators . Spring Loaded HingesPLF Half Shell HSSVSSPneumatic Actuators Jettison Spring ForceCoGCoG 1) HSS & VSS Separation 2) Rotation 3) Disengagement and Jettison

Document

AMS_2020.pdf

346 Actuator Selection : The key selection criteria to fulfill best the system goals are: • Aiming at a cost -efficient solution, the actuators are modified commercial of the shelf (COTS) products. • The system should be usable in a broad range of launcher sizes and separation conditions . Pneumatic actuators give the advantage to tailor its performance by the adaptable energy storage via the pressure. T his allow s a mission customized, optimal rotation initiation. For this , the very same qualified actuator is compatible for a wide range of applications. • To lower the shock coming from the separation process, the work to s eparate the PLFs is aimed to be applied with low force over a long stroke which leads to a very smooth transformation to the needed kinetic energy for the required movement . Gas springs provide a force over the full (adaptable) stroke, where the force reduction over stroke is depending on the (adaptable) gas volume ratio from contracted to extended configuration. This brings also a remarkable advantage in comparison to commercial steel springs where a long stroke providing continuous high forces leads to an increase in total length and with this a need for increase d spring diameter or additional suspension structure due to potential instability. • The gas spring stores the energy for the movement within the actuator itself . This has an advantage over other pneumatic actuators in terms of energy per weight ratio because no additional gas storage tank is needed. • A high qualified pressure capability of the actuator leads to a relative high energy per weight ratio that can be achieved. • Gas springs of this type are normally used in airplanes and trucks . Consequently, they have high heritage in demanding environment resistance and s olidity against vibration. • The preload ( pressurization) can be done in the very last moment before flight in the contracted configuration. All installation work can be done safely without internal forces. Given the goal for a scalable system and usage of the actuator on different launchers, together with the dependence of actuators location and launchers geometr y to the required actuators force and stroke, the gas springs showed during the design development the advantage of being very flexible in application. Design Description: In Figure 3 the actuator is shown as a schematic with explanations for the individual parts . Figure 3. COTS Gas Spring Actuator Schem atic The COTS actuator is composed with the following parts: 1. Cylinder contains the pressurized gas (N2) and guides the piston movement. 2. Piston rod is the moving part of the actuator and defines the stroke. Its diameter defines the active area the pressure acts on for the force. It is covered by a very strong protection against corrosion. 3. Connecting parts are adapted to meet the INTERFACE to the fairing . They are designed to sustain the vibration during launch, to transfer the force during rotation and enable the actuator to stay completely on the fairing after separation. 4. Piston bearing is the small part inside the cylinder that guides the rod and contains a jet , enabling the use of the full cylinder cavity as pressurized volume and allows the control of the movement velocity . 5. Guiding piece that seals the pressure volume and guides , together with the piston adapter, the movable piston rod. 1 2 34 5

Document

AMS_2020.pdf

347 An important difference of the selected gas spring versus other pneumatic actuators is that the full cylinder is used as pressurized volume. The total amount of energy, the piston force and the force reduction over the stroke can be adjusted towards the systems requirement by t he ratio between cylinder diameter, influencing the amount of energy, and piston diameter, influencing the force of the actuator . Depending on these parameters, a force reduction from beginning of stroke to end of stroke of less than 15% can be achieved without increasing the total size of the actuator. For pressurization a non- flammable, non- explosive and non- toxic gas is needed. For this the inert gas Nitrogen (N 2) is chosen, which meets all these requirements and is easy to purchase. In collaboration with the supplier , the actuator is tailored to the needs of the project. T hese adaptions include: • Material adaptions to meet ECSS standards2 as coating for the housing and sealing for bearing elements. • Construction of a pressure inlet valve adapter . The actuator is to be pressurized on launch si te, right before launch, enabling easy and save mounting after PLF closure. • Pressure sensor adapter in order to surveille the pressure processing. The sensor is mainly implemented for this project phase to verify the performance of the actuator. During the vibration and life testing the sensor provided important knowledge about the sealing qualities under external loads and the pressure behavior when exposed to temperature gradients and lifetime tests. As the sensor is also already qualified, the sensor can be used to monitor flight data or be omitted for standard implementation during flight to reduce the mass . Together with the supplier customization, the interface to the fairing was constructed. They are constructed to fulfill the following requirements: • The interface has to sustain the internal ly stored energy for the rotation during all phases before separation • All vibration loads during launch before separation need to be accommodated by the interface . • The interface needs to provide to the actuator the required degree of freedom to allow a smooth movement without jamming. • After PLF separation, the actuator is to be separated from the launcher and gets jettisoned with the PLF. • The concept needs to be adaptable to different launcher classes. The interface bracket s, designed to fulfill these requirements , comprise two different adapters for the PLF and launcher interface. Whereas on the PLF side an eye bearing provides the needed free moving space but close connection to hold the actuator at the payload shell, a ball joint on the launcher side prevents the actuator from jamming and releases the actuator after the stroke of the actuator has been reached. In Figure 4, the actuator with the interface bracket s for the implementation to a PLF is shown. In addition to the analysis of the bracket ’s load sustainability, an analysis has been run showing that the fairing can accommodate the actuator ’s loads during launch and separation. A focus has been on the fairing’s deformation caused by the actuator ’s force. It has been found that the deformation is negligible for the developed system as clearance towards possible fairing breathing is ensured during the rotation phase before jettison. The interface brackets have been designed and analysis for a medium launch vehicle. For the qualification campaign, described in the next section, the full set -up, including the interface joints, has been tested.

Document

AMS_2020.pdf

348 Figure 4: Actuator I/F schematic Characterization and Qualification: Using a gas spring as the actuator brings the difficulty of a changing pressure and resulting force with changing temperature. Pressurization needs to be adapted in a way to have the required force at separation, but the actuator needs to sustain a possible high temperature increas e. Thus , the characterization and qualification campaign enveloped a broad range of temperatures , for different launcher classes and different missions . This temperature range includ es possible low temperatures when the actuators are installed near a cryo stage of the launcher as well as the temperature increase caused by air friction during launch. The qualification campaign (including lifetime and thermal characterization) has been performed on different COTS actuator sizes, applying full mechanical and thermal loads as per medium launch vehicles PLF qualification requirements . As a result, a fully qualified pneumatic actuator at low mass and broad application possibilities has been found. Lesson Learned: One of the lessons from testing is that the material for the sealing is critical in terms of heat capability and sustainability to the vibrations loads during launch. Dedicated testing has been performed to select a suitable material enveloping the range of temperature, the range of pressure and the high vibration load requirements. Hinges The hinges guide the PLF during the rotation and provide the kick -off force to jettison the PLF halves. Two hinges are used per PLF half shell to ensure a balanced guidance of the PLF during rotation and enabling separation in the launcher’s rotational accelerated cases by avoiding high moments. The hinge’s main functions include the following: • Provide the mechanical interface during the fairing rotation phase. • Transform the actuators forces to rotational movement of the fairing. • Guide the fairing during rotation with low friction. • Accommodate external f orces during rotation. • Disconnect the mechanical connection at a given angle. • Provide a kick -off energy to jettison the PLF halves ensuring clearance. • Provide the correct force direction for the kick -off.

Document

AMS_2020.pdf

349 The developed solution is a very simple design to ensure passively the release at a given angle providing a kick- off energy to ensure clearance to the launcher. The hinge, in disassembled condition, consists of two parts as visible in Figure 8: • The launcher bracket is attached to the launchers last stage. It includes the bolt for the rotation axis and the protuberances that lock the hinge in closed configuration and counteract the k ick-offsprin g. This part is fixed on the launchers last stage and cannot be jettisoned with the fairing. • The hinge housing is mounted to the fairing halve- shell and is jettisoned with the PLF . It accommodated the kick -off-spring and the guiding pins, located in the slid of the launcher brackets protuberances in closed configuration. During launch, the hinge is in ‘closed’ condition as visible in Figure 8. The geometry provides the separation of the hinge parts and the release of the loaded kick -off-spring . The preloaded spring is accommodated within the hinge housing, pressing against the rotation axle bolt and counteracted by the launcher brackets protuberances. During the rotation phase, driven by the actuators, the hinge gets tilted into ‘open’ position. During this movement, the g uiding pins slides over the surface of the l auncher bracket’ s guiding protuberances. When the guiding pins reach the end of the launcher brackets protuberance the hinge is ‘open’ . The geometr ic constraint s are not given any more, the hinge housing is not locked to the launcher bracket anymore, and it gets pitchforked, with the fairing, by the released kick -off-spring system. Two of the described Hinge systems are installed at each PLF half shell. This avoids moments on the hinges, coming from external forces on the PLF. The alignment of the hinges needs t o be adjusted on each fairing type. This can be done in a simple way by building interface brackets to meet the launchers shape within the following geometrical constraints, mainly respective to the alignment of the fairing’s rotation axis during the separ ation event : • The hinges common rotation axis needs to be outside the launcher’s envelope as visible in Figure 5. This is necessary as to avoid jamming or the need of tr ansversal movement before the rotation phase. • The rotation axis needs to be below the fairing’s interface line to the launcher last stage, in order to avoid jamming at horizontal or vertical connection line. The correct placement is also dependent on the geometrical shape of the horizontal and vertical interface line. • The hinge kick -off-spring force direction is to be aligned with the fairing center of mass. This avoids additional moments introduced by the k ick-off and enables optimal use of the kick -off-energy for clearance. • • • Figure 5: Hinge alignment (1) In the following section, the development from concept to qualified system is shortly described including research studies and lessons learned.

Document

AMS_2020.pdf

350 Development : In a preliminary development study , the principle of the hinge was developed and a functional verification was run with a modifiable demonstrator model , shown in Figure 6, to prove the concept functional ity. The hinge for the demonstration model is designed to have a broad range of adjustable functional parameters as to accommodate different springs and featur ing an adaptable spring force and stroke length to vary the kick-off energy. The demo nstrator consists of a fairing dummy, down scaled from a large launch vehicle. The dummy was constructed with adaptable mass distribution, in order to simulate possible design changes on the fairing for the new separation system as well as asymmetric mass distribution to assess the robustness of the system by test . Only one hinge has been attached to the demonstrator, allowing to investigate and test occurrence of jamming due to asymmetric actuator force introduction. Figure 6: Early stage hinge during concept test ing on the demonstrator model The demonstrator test verified the principle of the system as well as the concept of the hinges. Several test variations in terms of initiated force direction, energy variation of actuator and kick -off-spring were conducted and showed the high robustness for the new jettison system. By varying a broad range of parameters, the concept also proved to be scalable to the use for different launcher classes with low additional work. Lesson Learned and Research: Along with the successful verification of the hinge’s principle and initial correlation to analysis, t hose tests showed that an improvement of the tribological behavior of the hinge’s gliding surfaces was necessary , i.e., the kick -off plunger to the axle and the suspension to the guiding pin. Consequently, a dedicated trade off and testing campaign was conducted to find a suitable material combination which ensures a low friction and avoids abrasion, jamming or cold weldin g. Three different material combinations , including a bearing, were selected by a tradeoff and tested extensive ly for gliding movement with high forces towards each other . The involved materials for the gliding surfaces are aluminum with Ematal coating and stainless steel. For the counter -acting surface t hree different materials and corresponding coatings were tested: • Material 1: Titanium with Dicronite coating • Material 2: Stainless steel with fiber reinforced plastic composite bearing • Material 3: Stainless s teel with Teflon- filled coating

Document

AMS_2020.pdf

351 A hinge dummy has been buil t simulating the internal loads by the preloaded kick -off-spring and external loads , simulating forces and moments on the hinge caused by the separation actuators and the launchers vertical and rotational acceleration at separation. In Figure 7 the results in terms of friction are presented. Figure 7: Material Test for the Gliding Surfaces; Test Set -up and Results Material 1 was discarded due to the high friction increase occurred at high loads . In contrast, both materials 2 and 3 achieved the test success criteria of low friction and wear. Comparing material 2 and 3 , combination 2 required additional parts (bearings), making the design more complex. In addition, material 3 also performed to high loads without change of the friction coefficient and no degradation was seen after exhaustive lifetime testing . As a result , the gliding surfaces material selection process was completed by the selection of material 3. In addition to the material selection, the interfacing geometry of the hinge dummy for the material tests were built similar to the qualification model hinges, including a preceding analysis for the hinge tolerances to ensure that no jamming occurs in the full temperature range due to different material expansion. This allowed preliminary functional tests to verify the concept against jamming in case of lateral or torsional external loads on the hinge dummy set-up, including life- time tests for the material and surfaces. Subsequently, dedicated flight model hinges were designed, built and tested to qualification loads according to the customer requirements for a middle launch vehicle (MLV) PLF. The campaign included thermal tests, vibration and lifetime . In Figure 8 the fully qualified hinge is shown, mounted on an MLV fairing with dedicated adapter brackets. The generic layout of the hinge allows the use on several different types of launchers . The only part needing additi onal design work are the INTERFACE adapter brackets to align with the fairings and upper stage curvature with the above described geometrical constraints . Additionally, the hinge is outlined to be scalable, the size is adaptable to accommodate forces from different launcher classes, if required. 1

Document

AMS_2020.pdf

352 Figure 8: Hinges for MLV PLFs Full Scale Test One of the advantages of the system is the possibility to test under ambient conditions without the need to use a vacuum chamber . To verify the functionality and correlation to the analysis , a full-scale separation test on an MLV fairing has been performed. Figure 9 shows the test set -up at RUAG Space premises in Zürich/Switzerland . For this test, the fully qualified mechanisms (hinges and actuators), from the qualification campaigns as described above have been used. Figure 9: Image of the PLF during rotation of the full-scale test For several separations runs, tracing the fairings movement with high -speed camera targets and accelerometers at different failure modes and degradation cases, the separation and jettison hinge and

Document

AMS_2020.pdf

353 actuator system have demonstrated to be a simple and robust system with repeatable and predictable results. Figure 10 shows the trajectory of the fairing during and after rotation and jettison for several test runs with varying energy. The picture shows the movement of the hinge in radial (y) and gravitational ( -x) direction, captured and tracked with a high -speed camera. The movement is robust against the variables that reduce performance of actuators and spring kick -off spring energy and can be adjusted very accurately. Figure 10: Full scale test trajectory Each run show ed similar rotation behavior , accelerations, trajectory and rotational speed. The test outcome comparison to FEM prediction results showed good correlation, considering the prediction of the effect of air resistance. Besides the repeatability and good FEM correlation of the rotation and jettison system, the methodology and procedure to test in ambient air by track ing the trajectory and accelerations has proven to be a reliable methodology to verify the system for future separation tests on different launchers . Conclusion This paper summarizes the activities carried out by RUAG Space within the Future Launchers Preparatory Programme – separation and jettison project. In the frame of this project, a low -shock jettison system has been developed based on the functional principle of a rotation, disengagement and jettison by means of passive gas springs actuators and spring loaded hinges. Q ualification models of the mechanisms have been built and qualificatio n testing has been successfully completed on component level and on system level in a full -scale test (TRL 6) . An extensive test campaign was carried out to verify the suitability and prove the advantages of the newly developed technologies with regards to simplicity and verification efforts. Additionally, a full -scale test has been successfully conducted for the validation of the separation analysis model proven functionality of the technologies developed. Overall, the test results are proving the functionality of the new jettison system. Acknowledgements The authors would like the thank the European Space Agency (ESA) and in particular the Future Launchers Preparatory Programme for supporting and granting a part of the investigation shown in this proceeding, in the frame of the Contract No. 4000121430- 17-F-JLV.

Document

AMS_2020.pdf

354 References 1. European Space Agency, 2012. ARIANE 5 fairing separation system undergoes testing. http://www.esa.int/spaceinimages/Images/2012/06/After_separation2 (accessed 28.01.2020) 2. ESA Standards: ECSS- Q-ST-70C rev2: Materials, mechanical parts and processes 3. Sanchez A., Halter B., Gerngross T., RUAG’s development of a modular low -shock jettison system, International Astronautic Conference IAC -19, October 2019, Washington, USA 4. Caldirola L., Schmid B., Cattaneo M., Schiaffini A., 2017. Test HDRM” development for separation performance verification.. In European Space Mechanisms and Tribology Symposium. 5. Arianespace, 2014. VEGA User's manual issue 4. Online, available in http://www.arianespace.com/wp- content/uploads/2015/09/Vega- Users -Manual_Issue- 04_April - 2014.pdf (accessed on the 04 .06.2019) .

Document

AMS_2020.pdf

355 Deployment Mechanism for an Earth Re- Entry Foldable Cloth Decelerator Carl E. Kruger* Abstract Adaptable Deployable Entry and Placement Technology (ADEPT) is a NASA technology development project that uses a folded, 3D woven carbon fabric decelerator to serve both as thermal protection and primary structure . A sounding rocket test flight, with a 3U Cubesat sized payload , was used to evaluat e ADEPT’s in-space deployment performance and supersonic st ability . Size, mass and complexity limits dictated a solely mechanical mechanism to deploy the folded cloth decelerator surface. This paper describes the design constraints , system design components, hardware development, testing, flight test execution, and flight test results of the deployment mechanism used by this “Nano- ADEPT” [1]. Introduction ADEPT is a technology development project that uses a foldable carbon fabric “skirt” to serve as the primary drag surface for an entry vehicle decelerator . The carbon fabric acts as an aerodynamic surface to provide drag while also being able to withstand the high temperatures of re- entry . Using a folding design allows a larger decelerator to be packaged with in any given launch vehicle ’s payload envelope. For small spacecraft missions, a small version of ADEPT (sub- 1m) could be used to provide a lower heating entry environment while still fitting within common secondary payload accommodati ons. This ADEPT Sounding Rocket One (SR -1) project focused on subsystem level testing of a 0.8-meter (NanoADEPT class) decelerator . Eight gores of a multi -layer, woven carbon fabric are sewn together with carbon thread to form a faceted 70° “cone shaped” aeroshell . A wind tunnel test (Fig. 1) was used to characterize the deflected fabric shape as a function of aero load, angle of attac k, and cloth tension . Arc jet testing (Fig. 2) exposed carbon cloth samples and seams to heating rates and heat loads expected for a Mars mission . Lastly , a sounding rocket test was executed to investigate spacecraft re- entry stability as the vehicle decelerated from Mach 2.8 to Mach 0.8. Figure 1. ADEPT Fabric Deflection Wind Tunnel Test Figure 2. ADEPT ArcJet Heating Test * NASA Ames Research Center, Moffett Field, CA Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 2020

Document

AMS_2020.pdf

356 Flight Concept of Operations The ADEPT SR -1 vehicle was designed for the following concept of operations: It would be launched to an altitude of approximately 100 km (62 mi) in the payload module of an UP Aerospace SpaceLoft XL sounding rocket . Shortly before apogee, ADEPT SR -1 would be ejected by a spring- loaded sabot assembly in a “tail first” orientation, so that as it fell to earth, it would re- enter nose first . Approximat ely 40 seconds after being ejected (to reduce risk of re- contact), the on- board timing circuit would command the carbon fabric decelerator to deploy . SR-1 would reach a peak Mach number of 3, at approximately 70 km (43 mi) altitude with no significant heat ing. Supersonic and subsonic stability would be determined using recorded accelerations, body rates, and orientation references for stability analysis at a later time. A hard impact of 25 m/s would be attenuated with crushable material . Data stored on hardened SD cards would be recovered from the impact site (no telemetry was used) . Design Constraints The deployment mechanism design was constrained by mass, volume, and C enter of Gravity (CG) limitations . An internal volume with dimensions of 10 cm x 10 cm x 30 cm (to simulate a possible 3U CubeSat payload) was reserved in the center of the vehicle (for this demonstration flight, this payload area housed batteries, electronics, sensors, and impact attenuation foam) . The external boundary was d efined by the 25.4- cm (10- in) internal diameter of the UP Aerospace launch vehicle payload module (Fig. 3). The deployment mechanism, including the cloth and ribs, was limited to the space between these two volumes . The overall length of the vehicle was constrained by the need to avoid high-speed flow impingement on the aft payload as the wake expands off the aeroshell shoulder. Figure 3. ADEPT SR -1 Prototype within Simulated Payload Volume Cloth shape and mass properties were constrained. For aerodynamic purposes, cloth deflections were required to be limited, and similar to those characterized in the wind tunnel test . The ribs and struts had to lock into position when fully deployed so the cloth maintained a constant tension during re- entry . In order to maintain stability during flight, the vehicle’s CG location and M ass Moment of I nertia (MMoI) were closely tracked and adjusted. ADEPT generally employs a rigid nose cap, which is used to cover the central payload as well as the deployable aeroshell when it is in the stowed position. In a typical high- heating re- entry environment , such as for Venus , it would carry a traditional thermal protection system , however for SR -1 a metallic nose cap was sufficient . As the carbon cloth skirt transitions from a stowed config uration to a fully open shape, the cloth must slide underneath the nose, and the nose cap must contact the cloth securely at the end of travel to prevent gaps that could allow hot gases to enter the payload region.

Document

AMS_2020.pdf

357 Figure 4. ADEPT SR -1 Flight Unit in Deployed Shape Design O verview ADEPT SR -1 uses coil springs to deploy the ribs and fabric . Long extension springs provide most of the deployment with a small force and long displacement . A retention cord constrains the rib tips . It is severed after launch and a moving ring which pushes the ribs outward is pulled towards the nose of ADEPT . Near the end of travel, a trip mechanism releases plungers containing short , stiff compr ession springs, which provide the final high force small displacement latching. In addition, the nose cap clamps down on the fabric (which needs to float as the ribs are deploying) just before latching . The time between retention cord severing and fully -locked deployment is less than one second. Figure 5. ADEPT SR -1 Layout

Document

AMS_2020.pdf

358 Framework The basic mechanism works like an umbrella, with rigid ribs and struts . The carbon cloth is attached to the eight ribbed structure resembling an umbrella frame. The cloth is secured at the tips of the ribs and the cloth rests on the ribs . It is not attached at the center, so the tension in the cloth distributes evenly over the eight gores as the ribs are deployed. Figure 6. ADEPT Prototype Showing Deployment A moving ring containing the anchorage for the eight struts slides along the axis of the vehicle on four linear slides . As the ring moves from the tail to ward the nose, the struts push the ribs out to a 70- degree angle. Unlike a conventional umbrella where the struts meet at a small ring, the SR-1 design employs a much larger square- shaped moving ring which encircles the 3U Cubesat payload volume. The ribs pivot from the nose of the spacecraft and are pushed outwards by struts located 1/3 of the way along th e ribs (Fig. 7) . Figure 7. ADEPT SR -1 Ribs and Struts

Document

AMS_2020.pdf

359 Figure 8. ADEPT SR -1 Centerbody with Moving Ring and Nose Cap Cloth Considerations The three- dimensional woven carbon cloth used on SR -1 is not strictly part of the deployment mechanism, but its properties are important to understand. The carbon cloth used for this version of ADEPT is a four - layer version, which demonstrated suitable folding behavior and tensioning ability (the cloth is st iff and becomes too thick to fold in 6- and 8- layer weaves) . Eight triangular shaped gores are stitched together using carbon thread to form an eight sided “pyramid” . A carbon cord is threaded through pockets at the trailing edge, to provide a defined trailing edge shape and tension. As mentioned earlier, t he clot h is free to float at the nose (Fig. 9 ). It is clamped at the underside of each rib tip (Fig. 10). Figure 9: ADEPT Floating Cloth under Nose Cap Figure 10: ADEPT Rib Tips with Kickoff Springs Rib retention and release When stowed, the eight ribs are parallel with the body axis, so there is no mechanical advantage for the long travel springs at the beginning of travel . For this reason, spring steel “kickoff springs” we re added to each rib tip to initiate rib movement . These push the rib tips away from the centerbody far enough to initiate rib movement . The rib tips are sculpted to guide the cloth into a trailing edge geometry that matches the

Document

AMS_2020.pdf

360 projects’ CFD model . In order to increase the vehicle’s Ixx (rotational inertia about the central axis) to facilitate spin stabilization, t he rib tips are stainless steel . ADEPT SR-1 is secured in the stowed condit ion by looping Vectran cord through hooks on all eight rib tips (Fig. 11). The Vectran is threaded through a pair of redundant spring- loaded Nichrome hot wire cord cutters [2]. Figure 11: Vectran C ord Loop and Nichrome Burn Wire Assembly Motive force springs The project decided against using a motor and gearbox due to packaging, complexity and power budget reasons . Precluding an electric gearmotor to drive the umbrella open led to the use of springs as the deployment motive force. However, meeting the cloth tension requirement of 17.5 N/cm (10 lbf/in) derived from the wind tunnel testing required high force as the cloth was close to full deployment . This is contrary to the properties of extension and compression springs, which have their highest force at their largest displacement and a much smaller force when approaching their free height. The mechanism designed to solve this conundrum uses two sets of springs to a ffect the required displacement and large final force. Eight pairs of long extension springs are used to provide the gross movement of the moving ring. A second set of stiff compression springs are triggered near the end of the range of motion, to creat e the necessary high force to tension the cloth and lock the ribs into position . The stiff compression springs also provide the motion to pull the nose cap firmly against the cloth and clamp it at full deployment . Figure 12. ADEPT Prototype showing Long- Travel Extension Springs

Document

AMS_2020.pdf

361 Figure 12 shows the long travel extension springs . They are stretched between the forward end of the spacecraft and the moving ring. There are two extension springs at each location, one nested within the other (for a total of 16 tension springs) , to provide suffici ent force without excessive free length. Moving ring The moving ring (Fig. 1 3) is an assembly of four identical machined quadrants , guided on four small linear slides, and fits in the space between the payload and the cloth. The stiff compression springs are housed in plungers that slide within the moving ring. The moving ring is aluminum and the plungers are stainless steel . No bushi ng or lubricant is used. Figure 13. ADEPT Moving Ring Assemblies (both sides shown) Nose Cap and Pins The nose cap (Fig. 1 4) was machined from stainless steel, to shift the CG forward . The cap contains four pins that are aligned with the plungers that are on the moving ring . The nose and pins slide in flange mounted linear bearings that are mounted to the forward end of the ADEPT centerbody (Fig. 1 5). The nose cap has approximately 4 mm of linear motion. It is lightly sprung away from the centerbody to allow for the cloth to move underneath the edges of the cap. The deployment mechanism pulls the nose cap down tight against the cloth after the system has reached full deployment. Figure 14. ADEPT Nose Cap with Pins and Figure 15. ADEPT Centerbody Forward Structure with Proximity Switches Rib Pivots and Nose Pin Bushings

Document

AMS_2020.pdf

362 Figure 16 shows the changes in section of the aluminum pins. The pins are the “triggers” for the multiple functions that occur during deployment. At the tip of the pin is a point, then a relief, and at the shoulder of the pin is a taper ed ramp with a recess on the back side . As the moving ring slides from the tail of ADEPT towards the nose, the pins pass through the center of the plungers . Figure 16. ADEPT Nose Pins and Latches Spring Plungers and Latches One end of each compression spring is seated against the base of the hollow plunger, and the other end is constrained by a tabbed washer that fits through slots in the plunger (Fig. 1 7). The ID of the compression spring is large enough to clear the pin. Figure 17. ADEPT Spring -Loaded Plunger (Compressed State)

Document

AMS_2020.pdf

363 This tabbed washer also limits the plunger rotation and displacement. As the pins pass through the spring plungers, the pointed pin tips spread t wo sliding, sprung plate catches (Figs. 18, 19) at the ends of the plungers . The plate catch es capture the tips of the pins by engaging in the relief . As the moving ring continues towards the nose of SR -1, eventually the taper ed ramp on the pins spreads a pair of l atches that have held the spring -loaded plungers in a compressed position (Fig. 20) . Figure 18. ADEPT Nose Pin and Plunger Catches Figure 19. ADEPT Nose Pin Sliding Plate Catc h Figure 20. ADEPT Plunger Assembly with Notch for L atch The four plungers extend, and since they have captured the tips of the pins, the nose cap is p ulled aft . This causes the now -deployed cloth to be clamped at the nose, which was a requirement . Additionally, once the nose cap reaches its end of travel, any further movement of the pl unger will cause the moving ring to be forced towards the nose of SR-1. This is the last bit of high- force movement required to pull the cloth to full tension and latch the deployment mechanism.

Document

AMS_2020.pdf

364 Figure 21. ADEPT Moving Ring Quadrant - Plunger in Latched (left) and Triggered (right) States Once the plungers have reached their full end of travel, the two latches that released the plungers are pulled back against the pins and capture the backside of the large diameter tapered ram p section of the pins (Fig. 22). This is the primary “locking” mechanism that prevents the moving ring from moving under aero load (the secondary locking mechanism is when the tips of the nose cap pins are captured). Figure 22: ADEPT Plunger Assembly, Latched in Fully Deployed Position

Document

AMS_2020.pdf

365 Figure 23: Fully D eployed ADEPT Showing Extended P lungers with Captured Pin T ips Instrumentation SR-1 contains a basic flight computer with an Inertia Measurement Unit and a Global Positioning System (GPS) to measure and record the vehicle’s accelerations, spin rates and position vs. time. In addition, a C - band transponder and antenna allow for range tracking, and a GPS beacon help s aid recovery . A Go -Pro camera is aimed at the trailing edge of the cloth skirt to observe any flutter if it occurs and records a set of status LEDs (Fig. 24). The deployment mechanism utilizes two switches and a burn wire. To sense ejection from the spacecraft, a pair of proximity switches are mounted in the nose cap. The corresponding trigger magnets are mounted in the sabot that cradles the nose cap within the payload volume. Once SR -1 is pushed free of the payl oad module, the switches change state and start a timer to release the ribs . After 40 seconds (sufficient time to guarantee no recontact with the payload module or nose cone), the burn wire severs the Vectra n cord and the mechanism deploys . At 69° rib angle (out of 70° total), a bypass style micro switch is tripped by the moving ring, which signifies “full deployment” . The separation, burn wire power, and “fully deployed and latched” events are indicated by LED’s that are visible in the field of view of the camera (Fig. 24). Figure 24: View of Trailing Edge of ADEPT Carbon Skirt after Exo- Atmospheric Deployment

Document

AMS_2020.pdf

366 Testing This dual spring mechanism design was extensively tested thru multiple prototypes, from a single plunger of printed plastic, through a 360- degree demonstration unit, culminating in the fabrication and testing of two flight units . The nested pairs of long travel extension springs were vibration tested at the component level, in the ir extended position, to check for possible resonance at launch vibr ation levels . The entire SR -1 was also vibration tested after electronics installation. Sine sweeps and three axis random vibration tests were performed on the stowed SR -1 to simulate the launch environment, followed by a full functional check. Functional testing of the mechanism (separate from integrated testing after the electronics were installed) included: 1) full deployment within 10 seconds of restraint release, 2) verification of full deployment via rib angle measurement, 3) verification of deployme nt diameter of 0.7m, 4) measure ment of gore deflections under known pressure loads using a vacuum bag (Fig. 25 ), 5) full deployment confirmation via limit switch, and 6) plunger assemblies latched/locked after deployment . Figure 25: Cloth Deflection Measurement Test using Vacuum Bag During a simulated payload ejection, the onboard Inertia Measurement Unit was used to estimate SR-1’s tip-off rates with respect to the payload module. Both stowed and deployed configurations were tested for CG location and radial offset, plus Mass Moments of Inertia . The soft goods presented a significant challenge in that the cloth is stiff and develops a “memory” after prolonged storage . ADEPT was shipped to the launch provider months in advance of the SR -1 flight, so it was important to be sure the cloth would unfold after long storage. The damping effect of the stiff cloth forced the cloth tension to be at the lower end of the allowable range. Adjusting for maximum cloth tension sometimes prevented the mechan ism from reaching full travel and latching, so a balance was required to meet deflection limits while achieving full deployment . Long- storage tests of 85 and 90 days were used to ensure that SR -1 would still deploy. Other than some non-deployments due to long storage at high cloth tension targets, t he deployment mechanism went through hundreds of cycles without other issues . To prevent possible contamination of the carbon cloth, initially no lubricant was used on the rib and strut pi vots. No lubricant was used on the plunger or catch assemblies. The linear slides and linear bushings were pre- lubricated by the manufacturer.

Document

AMS_2020.pdf

367 Lessons Learned Many design decisions were limited by the project budget . Pivots were simply aluminum holes rotating on steel shoulder bolts . The selected springs were chosen from commercial, off the shelf hardware, rather than custom made for optimal length/spring rate. There were no binding issues associated with the SR -1 shortcut of nesting the two long- trave l extension springs . One of the spring steel kickoff springs broke during testing and some lost their shape. The heat treat shop had trouble reaching 48- 50 Rockwell hardness, and suspected the spr ing steel was 1075 and not 1095 as indicated on the material certifications. An alternative to a flexible spring steel design, or more rigorous material certification , could prevent this . The carbon fabric was prone to unravelling at the edges of the gores , since it has no edge treatment . Carbon fibers would occasionally get caught in the linkage joints (Fig. 26). This was mitigated as best as possible by trimming any stray fibers before each stow. Figure 26: Carbon Cloth Fibers Trapped in Joint Figure 27: Wear at Pivot Joints SR-1 easily met its mass target of 25 kg (payload mass target of 15 kg) . The entire spacecraft weighed 10.9 kg ( 24 lb) . 0.61 kg (1.34 lb) of tungsten ballast was required at the front of the payload volume to meet axial (forward) CG location requirements . There is plenty of opportunity to wring weight from the design, especially the ribs, rib tips, and moving ring. Additional antennae and connectors were required aft of the burn wire assembly, which made the repetitive threading of the Vectra n cord (required during testing) difficult . Aft mounted antennae or interfaces may be a regular necessity . Design for accessibility would be helpful because of the need for repeated system testing. After repeated deployments during testing (50+), the mechanism began to fail to reach full deployment and latch . A small amount of Polytetrafluoroethylene (PTFE) grease was applied to the rib and strut joints, which helped (Fig. 27). For more repeatable deployment, the extra cost and complexity of bushings at the pivot joints and plungers would be useful. The Nichrome burn wire assemblies performed flawlessly . This very simple system was well suited for the SR-1 application .

Document

AMS_2020.pdf

368 Conclusions ADEPT was launched from White Sands Missile Range on September 12, 2018. It was recovered after its planned hard landing and inspection showed that the decelerator deployed and fully latched (Fig. 28). The on-board video recovered from SR -1 showed a fully deployed shape and an illuminated deployment indicator LED. Analysis of SR -1’s flight data showed angle of attack oscillations, but that it was stable from Mach 2.8 to Mach 0.8. Figure 28: Post -flight ADEPT Recovery Of the ADEPT SR -1 Success Criteria related to the deployme nt mechanism, Criteria D was defined as : “ADEPT achieves fully deployed and locked configuration prior to reaching 80 km (49.7 mi) altitude on descent. ” This was verified by evidence from the onboard GoPro video. Of the ADEPT SR -1 Key Performance Parameters related to the deployment mechanism, P arameter 1 called for : “Exo-atmospheric deployment to an entry configuration of the 1m- class ADEPT. ” This was met with a confirmed, fully locked deployment to 70 ° rib cone angle prior to entry, as evidenced by the onboard GoPro video, and post -recovery inspection. This mechanism proved successful and was well suited for our specific ADEPT case: an Earth re- entry aero- stability demonstration of a small diameter vehicle, with minimal heating and a low cloth tension require ment . Larger applications of the ADEPT concept, or entry environments with high heating or dynamic pressure, may require a motorized deployment design. References 1. Smith, B., Cassell, A., Kruger, C., Venkatapathy, E., Kazemba, C., Simonis , K. “Nano- ADEPT: An Entry System for Secondary Payloads .” IEEE Aerospace Conference, June 2015. 2. Thurn, Adam “Nichrome Burn Wire Release Mechanism for CubeSats .” Proceedings of the 41st Aerospace Mechanisms Symposium, May 16- 18, 2012.

Document

AMS_2020.pdf

© 2020 Lockheed Martin Corporation 369 Lessons Learned D uring the Development , Qualification , and Production of the MM S olar Array Thomas B. Pace* Abstract The Multi -mission Modular Solar Array (MM Solar Array) is Lockheed Martin’s fourth generation flex array. The MM Solar Array was designed to be a high- powered solar array , 7.5-12.5 kW and higher per Wing (15-25 kW and higher with 2 Wings per spacecraft ), that is easily configurable for a number of different missions. The prime objectives during the design process were a finished product that was cost -effectiv e, easy to manufacture , and modular. Like previous generations of Lockheed Martin flex arrays, the MM Solar Array Z -folds the blanket and stows it between two rectangular composite structures that protect the blanket during ascent on the launch vehicle. Once on orbit, the MM Solar Array deploys in two phases. During phase one of the deployment, the launch locks are released, and hinges move the MM Solar Array away from the vehicle and position the blanket for deployment. During phase two of the deployment, motorized actuators are energized and the mast is deployed, which in turns pulls the Z -folded blanket open. At the end of the deployment, two blanket tensioning mechanisms are engaged. These mechanisms ensure the blanket behaves as a planar membrane throughout the life of the vehicle, thus cr eating a stiff structure that can survive vehicle engine and thruster firings. Once deployed, the MM Solar Array measures approximately 23 meters (75 feet) from the base of the solar array to the tip. O ne of the bigger design challenges with the MM Solar Array was to survive an orbit -raising maneuver , i.e., an apogee engine burn , with two 23 -meter ( 75-foot) long MM Solar Arrays fully deployed on opposite sides of the vehicle. The structure and the hinges of the MM Solar Array had to be designed to survive this high loading condition. During the development, qualification, and production of the MM Solar Array many lessons were learned. In composite manufacturing, a critical lesson regarding cure profiles and how small variations can adversely change the behavior of a composite structure were learned. In mechanisms, proper handling of springdriven devices so as not to damage the device were learned. In test, a simple but important lesson was learned on est ablishing the correct test temperatures for a thermal cycle test of a complicated mechanical assembly. Also in test, the degradation of the strength of composite structures at cold temperatures was investigated. In the supply chain realm, lessons were learned about how one supplier might interpret what was thought to be a clear requirement completely differently than another supplier . During failure investigations, an important lesson was learned about obtain ing an independent perspective of the failure at hand. Some positive lessons were also learned, such as getting manufacturing and tooling involved early in the design process led to a smooth manufacturing process. Another positive lesson was that requalificati on of a heritage mechanism is a good idea, especially if the mechanism has not been built in over 15 years. The lessons learned in the development, qualification, and production of the MM Solar Array resulted in a cost-efficient and manufacturable solar array design that can be easily configured for any number of highpowered missions. Four MM Solar Array Wing Assemblies are currently in- orbit on two different vehicles. The launch of Wings 5 & 6 is scheduled for early 2020 . * Lockheed Martin Space, Sunnyvale, CA; tom.b.pace@lmco.com Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 2020

Document

AMS_2020.pdf

© 2020 Lockheed Martin Corporation 370 Figure 1 - Lockheed Martin Heritage Flex Arrays Figure 2 - Lockheed Martin MM Solar Array Introduction The idea for the MM Solar Array was conceived in 2013 when Lockheed Martin desired a high- powered solar array for its LM 2100 ™ fleet of commercial spacecraft. The goal of 25 kW per vehicle (or 12.5 kW per Wing) was established. The design team was faced with a significant decision: expand the current generation of rigid solar arr ays or utilize Lockheed Martin’s heritage in flex arrays to develop a low -cost, low-mass fourth- generation flex array. The trade studies that followed quickly concluded that the flex array solution had significant mass and cost benefits at power levels gre ater than ~15 kW per spacecraft , so the decision was made to invest in the flex array technology. For the next 5 years, Lockheed Martin invested significant internal research and development dollars to sy stematically design, develop, and qualif y the MM Solar Array for flight programs. Every decision in the design process was asked 3 important questions: h ow does it impact cost, how does it impact manufacturing, and how does it impact modularity? The last of these questions regarding modularity perhaps shaped the design of the MM Solar Array more than anything else. Knowing that different satellites have different power requirements, the design team mandated that the finished product be easily configurable for different power levels. The result was a very mod ular solar array design. There are 375 documents that define the MM Solar Array. Of these 375 drawings, only 15 of them require modifications to change the solar array from a wing that produces 12.5 kW to a wing that produces 6 kW or 8 kW or 10 kW, etc. And most of these drawing m odifications are very basic changes that simply change the length of a certain assembly, like the mast or the blanket. This modularity enables the MM Solar Array to be a fully qualified, turn -key solution for most any program. This means that on day one of the program, the design of the solar array is already 96% complete. Non- recurring engineering efforts are almost non-existent, and production of the MM Solar Array can start almost immediately after contr act award. Overview The MM Solar Array is divided into 4 major subassemblies. The first is the Boom Assembly. The structure of the Boom Assembly consists of hollow composite tubes with bonded titanium end fittings that are connected to each other across hinge lines. The purpose of the Boom Assembly is to deploy the MM Solar Array far enough away from the vehicle so that the vehicle does not cast a shadow on the solar array and to improve the thermal view factor for the spacecraft . Harnesses that transfer power from the solar array to the vehicle are routed along the Boom Assembly. The second major subassembly is the Deployer Assembly. The structure of the Deployer Assembly is made up of composite panels with aluminum honeycomb core and carbon- fiber facesheets . These panels are bolted together and support the mechanisms that are used to deploy the solar array blanket. The Deployer Assembly is a one- time use mechanism. Once the MM Solar Array is deployed, the Deployer Assembly serves as a piece of structure

Document

AMS_2020.pdf

© 2020 Lockheed Martin Corporation 371 that supports the overall solar array. The third major subassembly is the Blanket Container Assembly. Like the Deployer Assembly, the Blanket Container Assembly utilizes composite panels with aluminum honeycomb core and carbon- fiber facesheets. The purpose of the Blanket Container Assembly is to protect the Blanket Assembly during launch, position the Blanket Assembly for deployment once on orbit, and provide tension to the blanket during its on- orbit life. The fourth major subassembly is the Blanket Assembl y. The structure of the Blanket Assembly consists of a thin fil m of Kapton. The solar cells are bonded to the Kapton in segments called panels. Panels are then connected to each other via hinge pins and Z-folded in the Blanket Container Assembly to create the Blanket Assembly. Deployment The MM Solar Array deploys in two phases. The first phase of the deployment starts when the low -shock non-pyro release mechanisms that hold down the solar array to the vehicle are released. Once these 8 release devices are fired, 3 hinge lines deploy the Boom Assembly and Deployer Assembly. A f ourth hinge line rotates the Blanket Container Assembly and positions it for deployment of the Blanket Assembly. The end of phase one of the deployment is achieved when all four hinge lines lock out. Figure 3 – Stowed Configuration Figure 4 – First Stage Deployment Phase two of the MM Solar Array deployment consists of deploying the Blanket Assembly. Like previous generations of Lockheed Martin flex arrays, the Blanket Assembly is deployed usi ng a central backbone called a mast. The mast is a deployable composite structure that is connected to the outermost panel of the Blanket Assembly. When power is applied to the actuators, the mast is deployed and as it deploys, it pulls the z -folded blanket open. At the end of the mast’s travel, tensioning devices called the Tension Mechanisms are engaged. The Tension Mechanism is a constant force negator spring device that is mounted to the static structure of the MM Solar Array. A cable is wrapped around a central spool of each Tension Mechanism and then connected to the most inboard panel of the Blanket Assembly. As the mast nears its end of travel, the Tension Mechanism pays out a short length of this cable, thus applying a known tension to the blanket. The Tension Mechanism transforms the Blanket Assembly into a tensioned membrane, which is imp ortant to the satellite’s guidance and control because if the blanket is not tensioned,

Document

AMS_2020.pdf

© 2020 Lockheed Martin Corporation 372 then control of the satellite is compromised as it is very difficult to operate a vehicle that has two 23 -meter (75-foot) long “floppy” appendages. For the remainder of the mission, the Tension Mechanism is left in this “partially deployed” state, where it can pay out more cable or retract some cable while maintaining a constant force. This is important because the Blanket Assembly will expand and contract as the temperature changes, especially as the vehicle goes into and out of the Earth’s eclipse. Figure 5 – Fully Deployed MM Array in the Background, Stowed MM Array in the Foreground

Document

AMS_2020.pdf

© 2020 Lockheed Martin Corporation 373 Figure 6 - Stowed MM Solar Array Figure 7 - First Stage Deployment : Four Hinge Lines Deploy Boom Assembly, Deployer Assembly, and Blanket Container Assembly Figure 8 - First Stage Deployment Complete

Document

AMS_2020.pdf

© 2020 Lockheed Martin Corporation 374 Figure 9 - Start of Blanket Deployment Figure 10 - Second Stage Deployment Complete Orbit Raising Lockheed Martin’s commercial satellite heritage has been to deploy the solar arrays once all orbit raising maneuvers have been completed. With rigid solar arrays, power generation during this orbit raising phase is possible because when stowed, rigid solar arrays still have solar cells that can be illuminated by the Sun (the entire outboard panel can generate power during orbit -raising). However, the MM Solar Array does not have any exposed solar cells when stowed. This means that the MM Solar Array must be deployed as soon as separation from the launch vehicle occurs. The result is that vehicle orbit raising maneuvers must be done with 23-meter (75- foot) long deployed solar arrays on either side of the vehicle. The vehicle’s orbit raising engine drove signi ficant stiffness and moment -loading design requirements into the Boom Assembly and hinges , but the highest load was seen prior to the last engine firing when the vehicle was at its lightest mass.

Document

AMS_2020.pdf

© 2020 Lockheed Martin Corporation 375 Figure 11 - Orbit Raising with Deployed MM Solar Arrays Lessons Learned Overview During the design, development, qualification, and production of the MM Solar Array, many valuable lessons were learned. Lessons learned were plentiful during development and qualification, but some of the most important lessons were learned during the production of the flight MM Solar Array s. Some lessons were still being learned during the production of the fifth and sixth MM Solar Arrays. Thermal Testing of Complex Assemblies The Deployer Assembly is a complex subassembly of the MM Solar Array. It contains structural panels made from carbon- fiber facesheets with aluminum honeycomb core, precision mechanisms such as the drive actuators , large aluminum sleeves that guide the mast during deployment , and the mast itself. The mast of the MM Solar Array is a composite lenticular strut that is based on work originally done by Dr. George Herzl , a co- founder of the Aerospace Mechanisms Symposia. When deployed the mast is in a stress- free state. But when stowed, i.e. , flattened and w rapped up around a drum, the mast is in its most stressed state . The Mast is in its stowed state during thermal cycling of the Deployer Assembly. Before starting the qualification thermal cycle test of the Deployer Assembly, the hot and cold temperature extremes needed to be established for the test. A thermal model was used to determine the temperature extremes that would be experienced by each part of the Deployer Assembly during the mission. The analysis showed that the structur al panels of the Deployer Assembly were the driving components, so t he thermal extremes of the Deployer Assembly thermal cycle test were established based on these p anel temperatures. Once this assessment was complete, it was noted that t he drive actuators would have to be removed from this test because they contained circuit cards with electronic components that could not survive the planned temperature extremes (during flight, heaters are used to maintain the temperature of these circuit cards) . However, the team forgot to assess whether any other parts of the Deployer Assembly had any temperature restrictions. Instead, the team went forward with the qualification thermal cycle test and every remaining component in the Deployer Assembly w as tested to the predicted temperature extremes of the structural panels . Upon completion of the test, the mast was unrolled and deformities in the mast were observed. Further investigation revealed that the deformities were caused by subjecting the stowed mast, i.e. , the stressed structure, to a hot temperature extreme that was above the glass transition temperature of the composite resin . Certain members of the team were aware of this limitation, but this was not communicated to the team member that wrote the test plan. There are multiple lessons learned in this example. The first is that a stressed composite structure should not be thermal cycled above its glass transition temperature. Two is that the limitations of all components in complex assemblies must be known prior to beginning a thermal cycle test or any other test . And three is that all members of the team should be involved in establishing the thermal extremes for a thermal cycle test of a complex assembly .

Document

AMS_2020.pdf

© 2020 Lockheed Martin Corporation 376 Cure Profile of Composites Another lesson learned involving the mast revolved around the cure profile of the mast. Masts had been successfully produced for the development phase, the qualification phase, and even the production phase for two flight programs. But when production began on masts for the third flight program, a change was made that had dramatic unintended consequences. In an effort to be more af fordable, the production team changed the cure profile of the mast so the mast could be cured in a shorter amount of time. The change consisted of a quicker ramp rate to the cure temperature and a lower cure temperature. Both changes were acceptable per the tolerances on the engineering drawing. However, the tolerances on the drawing were quite large ( 20°C) and no one had ever built a mast to these parameters before . The result was a brittle mast that cracked when stowed, though the cracks did not present t hemselves until the third or fourth stow/deploy cycles. By that time, the masts had been fully integrated into the Deployer Assemblies before it was realized that they were discrepant and would have to be replaced. Cost and schedule impacts resulted, along with a lengthy failure review board investigation. The lessons learned in this instance are many. First, cure profiles of composite structures should not be changed after they have been proven to produce acceptable flight hardware. Or if they are changed, then adequate delta- qualification tests should be conducted to ensure the end product is still acceptable. Second, clear communication channels between production teams and engineering teams are essential . The production team made the change to the cure profile without any buy -in from the engineering team . The engineering team would have denied the request for the cure profile change if they had been consulted because they knew the cure profile was critical to this type of composite structure. Third, the engineering team needs to ensure that whatever tolerances are established for the cure profile are qualified. The engineering team assumed that the end product would be the same given one ramp rate versus another or given one cure temperature versus another. But the engineering team never validated this assumption through tests. If the engineering team specifies a wide range of ramp rates and cure temperatures, then the engineering team needs to ensure those ranges produce acceptable hardware. Figure 12 – Caution: even though these two composite cure profiles are within the tolerances, one may produce acceptable hardware while the other may not

Document

AMS_2020.pdf

© 2020 Lockheed Martin Corporation 377 Hinge Springs The next lesson learned involved the springs used in the hinges. During the first phase of deployment of the MM Solar Array, four hinge lines are deployed to position the MM Solar Array for deployment of the Blanket Assembly. During qualificat ion of the hinges, the springs used in the hinges were procured from a supplier that specialized in producing springs. But for the flight units, the procurement team changed the supplier of the springs to a general machine shop to save cost. This machine s hop delivered discrepant springs that were fabricated from the wrong material. But the discrepancy was not realized until after two sets of hinges had been built, tested, and installed on the flight MM Solar Arrays. Major cost and schedule impacts resulted and several work -arounds had to be implemented to meet critical vehicle need dates. Some basic but important lessons were learned from this event. First, communication between the engineering team and the procurement teams is important. The responsible engineer was not consulted on the supplier change for the springs. If he had been, he would not have approved the change. Second, if it is important that a product like a spring be procured from a specific source, then engineering should create a source- control drawing that specifies the approved suppliers. Hinge Studs Another lesson learned involving hinges revolved around a part of the hinge called the stud. This small machined part contains a threaded end with an undercut feature. The callout on the undercut was “R .015 undercut to minor dia. of thd”, i.e., undercut w ith a radius of 0.015 inch (0.38 mm) to the minor diameter of the thread. Even though this part and parts like it had been successfully fabricated for over 20 years, a supplier used to produce studs for the MM Solar Array hinges misinterpreted this undercut callout. This supplier interpreted the callout to be undercut 0.015 inch (0.38 mm) from the minor diameter. The result was a neck on the part that was too small (see Figure 9 ). During assembly of one of the hinges, the threaded part snapped off when a nut was torqued to it. This discrepancy was noted after multiple sets of hinges had been built, tested, and integrated into flight MM Solar Arrays. Major cost and schedule impacts resulted and several work -arounds had to be implemented to meet critical vehicle need dates. In the end, the flight drawing was updated to explicitly call- out the minor thread dimension and show the undercut to that dimension. The lesson learned here is that even if the part has been successfully produced by multiple suppliers for over 20 years, design engineers should always assess their drawings for any ambiguity because i f there is any ambiguity in the drawings, then someone may misinterpret it. Figure 13 - Correct Hinge Stud Undercut Versus Incorrect Undercut

Document

AMS_2020.pdf

© 2020 Lockheed Martin Corporation 378 Figure 14 - Old Hinge Stud Undercut Callout Versus New Callout (dimensions in inches) Tension Mechanism The next set of lessons learned involve the Tension Mechanism, which was discussed in some detail in the Deployment section. As discussed there, the Tension Mechanism contains a cable wrapped around a spool . One end of the cable is swaged to a threaded stu d while the other end of the cable is secured to the spool. The spool is connected to a constant force negator spring so as a result, the negator spring is constantly trying to retract the cable around the spool. During assembly of the Tension Mechanism and duri ng installation of the Tension Mechanism on the MM Solar Array, care must be taken to ensure that uncontrolled retractions of the cable do not occur. If uncontrolled retractions occur, then the threaded stud at the loose end of the cable can be uncontrollably slammed into parts of the Tension Mechanism and damage the mechanism. Damage was most often seen in the form of broken strands in the cable. In order to prevent these uncontrolled retractions, 3D -printed plastic parts were used as shop aids to restrain the motion of the cable. The lesson learned is that proper tooling is required to prevent unintended motion of spring- loaded mechanisms. One other valuable lesson learned regarding the Tension Mechanism involved the decision to build a dedicated qua lification unit. The Tension Mechanism design is largely unchanged from a heritage mechanism that has been performing nominally on- orbit for many years . Due to this successful heritage, the design team was tempted to not build a dedicated qual unit and jump straight into acceptance build and test. However, it was suggested to the team that since the product had not been built in over 15 years, it would be prudent to build a dedicated qualification unit. This decision turned out to be the right one because as requirements evolved, the design changed and deltaqualification tests were required to qualify these changes. The fact that a dedicated qualification unit existed made it easy to conduct these delta -qualification tests. Figure 15 - Tension Mechanism with Restraining Tool (red part)

Document

AMS_2020.pdf

© 2020 Lockheed Martin Corporation 379 Tooling One of the good lessons learned from the MM Solar Array experience was to get the production team and the tooling team involved early in the program. When these teams are involved early, they can get a headstart on what tools will be needed to most effectively assemble and test the product. By working alongside the design engineers, proper tooling can be designed and fabricated in time to support the flight builds. Often, a program may get to a certain part of the build and find that they forgot to build a critical tool or fixture, or a critical tool or fixture is not designed correctly. By engaging with the production and tooling teams early, the MM Solar Array team avoided this common pitfall. One of the most complex pieces of tooling was the system used to offload the MM Solar Array during deployment tests , called the Rail Offload System. The Rail Offload System had to meet several challenging requirements. It had to accommodate the entire 23- meter ( 75-foot) length of the MM Solar Array. It had to accommodate a lateral motion of about 3 meters ( 10 feet) to allow the hinges to open during the first stage deployment. It had to be as low friction as possible to simulate zero gravity. And it had to offload each major subassembly in the MM Solar Array. The Boom Assembly, Deployer Assembly, and Blanket Container Assembly were offloaded using one link to the Rail Offload System, but the Blanket Assembly required each of the 30 panels in the Blanket Assembly to be individually linked to the Rail Offload System. This was the biggest challenge because when the Blanket Assembly is in its stowed configuration, the Blanket Assembly is less than 2.54- cm (1-inch) thick. The challenge was to develop a solution to be able to package 30 panel offloads in such a small space. The solution had 3 major parts to it. One was that thin offload trolley s were used. Two was that two tracks were used, one to offload the odd- numbered panel s and one to offload the even- numbered panels. By doing this, the offload trolleys for adjacent panels could nest within each other, thus saving space in the stowed configuration. And three was that it turned out that the 30 offloads did not have to fit in a 2.54- cm (1-inch) wide space. It was acceptable for the offload cables to have a slight angle to them when the Blanket Assembly was stowed as long as there was some compliance built into the offload cable. Figure 16 - Blanket Assembly Rail Offload System

Document

AMS_2020.pdf

© 2020 Lockheed Martin Corporation 380 Moment Test at Cold Temperatures Long after all qualification tests had been completed and flight MM Solar Arrays were getting ready to be delivered to the vehicle, a question was asked regarding moment testing at cold temperatures. The concern was raised by an analyst who asked the team “I know you tested your composite boom tubes at ambient temperatures but how did they perform when they were tested at cold temperatures?” The team did not have an answer for this question. The team followed standard protocol for a composite boom tube, whi ch consisted of fabricating the tube, bonding in the titanium end fitting, thermal cycling the tube to the hottest and coldest temperature predicts plus margin, and then proof -loading each tube. However, the proof load test was always done at ambient temperature. No cold temperature moment testing was ever conducted, even on the qualification unit. But the design of the MM Solar Array relies on the strength of a single boom tube that connects the entire array to the vehicle. And this single tube sees a very high moment load at the time the vehicle’s liquid apogee engine (LAE) fires to raise the orbit of the vehicle (recall from an earlier section of this paper that the MM Solar Arrays are fully deployed prior to firing the LAE). And this tube could be cold a t the time the LAE fires. Knowing that there is some reduction in the strength of bond lines and composite structures at cold temperatures, there was suddenly an urgent need to fabricate a flight coupon and moment test it at cold temperatures. A coupon was fabricated and subjected to the appropriate moment load at temperatures as low as - 64˚C. Several different load cycles were conducted to simulate the different loads that the boom would experience as the LAE is fired at different times during orbit raising. The results were that the boom tube passed every test and the late- breaking concern was laid to rest . The lesson learned here is that critical bonded structural joints should be tested at temperature extremes to ensure they will survive this environment. However, the costs of such a test can be prohibitive so care needs to be taken when identifying critical joint s. A side lesson learned from this experience is that it is a good idea to include bolts in bonded joints where possible, a.k.a. chicken fasteners . Most composite tubes rely on the strength of the bond line that bonds in the end fi tting. But if bolts are also present in this joint, then the strength of the joint is increased. The MM Solar Array was originally designed with bolts in these bonded joints as an added measure of margin. One question that was never answered by the cold moment test is “what is the reduction in strength of a boom tube that is tested at ambient temperature versus a boom tube that is tested at cold temperature?” Is there a 10% reduction in strength? 25% reduction? 50% reduction? The team had hoped to answer this question by breaking a boom tube at ambient and then breaking one at cold. But due to budget constraints, there was only enough funding to build one tube and verify that it would survive the worst -case moment at the worst -case temperature, i.e. , a quali fication load of 618 Nm (5471 in- lb) at a qualification temperature of - 64˚C. Further research in this area would be prudent to determine the actual percent reduction in strength at cold temperatures. Figure 17 - Critical MM Solar Array Joint Figure 18 – Boom Tube Test Article

Document

AMS_2020.pdf

© 2020 Lockheed Martin Corporation 381 Obtain Independent Perspectives During the first deployment of the fifth flight MM Solar Array, an anomalous behavior was noted at the very end of the deployment. The behavior manifested as a sudden jerk of the blanket followed by ripples of the blanket along its length. The deployment successfully completed to lock -out, but this behavior was not nominal , so an investigation was started. The team investigated many possible causes and converged upon what they thought was the likely root cause, i.e. , that a bracket was undersized and not allowing a bar to be smoothly pulled out of this bracket. The team was in complete agreement that this was the likely root cause and was about to implement some trouble- shooting steps when it was suggested to the team that they solicit some outside input from a group of independent observers. It was not long before the group of independent observers were able to convince the team that an undersized bracket could not have caused this anomaly. Further inspection of photographs of the hardware revealed the true root cause, i.e. , the presence of a burr on the bracket. Inspection of the actual hardware confirmed the presence of the burr on this bracket and only this bracket. All other brackets had been properly deburred. During deployment, the burr behaved like a hook and momentarily grabbed onto the bar. As the force of the deploy ment overcame the burr, it suddenly released the bar, causing the sudden jerk in the deployment motion. The obvious lesson here is to ensure all burrs are removed from interfaces that move relative to one another. But a more important lesson is that schedu le pressures sometimes cause teams to want to quickly determine root cause and move on to the next step. If this team had done that, the evidence of the burr would have been destroyed and root cause may never have been determined. Teams must learn to take a breath and ensure that root cause has been conclusively determined before moving to the next step in the process. Another lesson learned is that it is always valuable to get an independent assessment of an anomaly from a trusted expert. The independent e xperts in this case were able to see the issue from a different perspective and challenge the assumptions that the team made to determine the wrong root cause. The independent experts were able to offer a more plausible root cause that upon further investi gation turned out to be the correct root cause. Figure 19 – Burr on Bracket

Document

AMS_2020.pdf

© 2020 Lockheed Martin Corporation 382 Conclusion Many lessons were learned during the development, qualification, and production of the MM Solar Array . The MM Solar Array has been successfully qualified to AIAA S-111 for Solar Cells and AIAA S-112 for Solar Panels, including continuous monitoring of the solar cells during the 1.5X thermal life cycle test. All mechanisms on the MM Solar Array have met the requirements of AIAA S-114 for Moving Mechanical Assemblies. Four Wings are currently flying on two commercial satellites that were launched in February and April of 2019 and two more Wings are set to fly on another commercial satellite in early 2020. As its name suggests, the Multi -mission Modular Solar Array stands ready to support multiple missions and is looking forward to powering the next generation of high -powered Lockheed Martin satellites. Figure 20 - MM Solar Array Integrated to Vehicle

Document

AMS_2020.pdf

383 Solid -State Hinge Mechanism for Simple Panel Deployment System Thomas (TJ) Rose*, William (Brad) Hensley * and William Francis * Abstract A common actuation method for flip-out panel deployment is a traditional torque hinge mechanism, which is successful largely due to its deterministic nature. However , these mechanisms typically require bulky support structures with high part counts. Another actuation method for flip -out deployments is a solid- state hinge mecha nism, or tape spring. Tape springs have been utilized for several satellite panel deployments in the past, where the primary advantages of these systems are simple design and low part count. With the space industry on the cusp of an industrialization surge , we have identified this integrated design to provide a distinct advantage over the complex traditional systems. This paper will discuss the development process for the flight -qualified Roccor Panel Deployment System ( ROC- PDS) with a specific focus on the analytical and empirical methods developed for solid- state hinge mechanism. Figure 1: Solid -State Hinge Mechanisms and Deployment system Introduction The ROC -PDS was developed at Roccor , from concept to flight qualification, as an integrated solution for utilization in large constellation applications. The defining advantage of this system is that the part count is an order of magnitude less than its torque hinge competitor. As shown in Figure 3, this system is comprised of a composite tube as the primary structure, with integrated solid- state hinges. By integrating the hinges in this way, the tube can take the place of the k ickoff spring, deployment torsional hinge, lockout mechanism, lockout damper, while still serving as the primary deployed structure supporting the pa nel. Development of the ROC -PDS required material and structural qualification methods developed by Roccor and the High Strain Composites ( HSC ) community. This effort included the development of qualification techniques for high strain composites, experi mental techniques for material performance measurement, testing and analysis techniques for the design and validation of subsystems like the solid- state hinges, as well as empirical and analytical methods for deployment kinematic predictions. New Space Qualification Challenges Roccor has successfully qualified and delivered HSC products for space- flight customers including 1) a radio- frequency precision furlable boom, and 2) a furlable antenna system; and is currently qualifying HSC products for space- flight customers including 3) an FCC -certified deorbit device, and 4) a solar array * Roccor, Longmont, CO Proceedings of the 45th Aerospace Mechanisms Symposium, NASA Johnson Space Center, 2020

Document

AMS_2020.pdf

384 deployment system. Each mission’s development and qualification program followed a similar test and analysis plan to that illustrated in Figure 2. Although these four missions were relatively risk tolerant, the tasks labeled with caution symbols in Figure 2 were very costly, schedule- intensive, and performancelimiting due to a lack of industry accepted practices applying to HSCs. The lack of standards creates a roadblock for the use of HSCs for high- value DoD or NASA missions , for which a new class of highperformance deployable structures would be enabling. Roccor views collaboration with the broader spaceflight community to develop industry -accepted thin- ply composite manufacturing methods, databases, engineering tools, test and qualification methods as vital to its business model. Figure 2. Roccor’s benchmark thin- ply composite development and qualification plan adapted from recent Roccor TPC space- flight programs (steps labeled with a caution symbol are costly, difficult and limiting due to the lack of applicable industry standards). ROC- PDS System Architecture As shown in Figure 1, the ROC -PDS is comprised of a panel, stowage containment fixtures , a hinged boom with end brackets, kinematic mounts, a hold down release mechanism, and two solar array drive actuators . The ROC -PDS was developed at Roccor from concept to flight qualification as a low part count solution for utilization in large constellation applications. The defining advantage of this system is that , by integrating solid-state hinges into the tube, the part count is an order of magnitude less than comparable mechanized systems. By integrating the hinges in this way, the tube can take the place of the k ickoff spring, deployment torsional hinge, lockout mechanism, lockout damper, and still serves as the primary deployed structure supporting the panel.

Document

AMS_2020.pdf

385 Figure 3: (Left) Stowed System configuration; (Center) Strain energy deployment simulation configuration; (Right) Deployed system simulation configuration Development of Solid -State Hinge Closed cross section tubes containing solid-state hinges have been demonstrated in the past for the use with several long antenna applications (Ref. 1). Much of the groundwork for this technology was developed by research facilities ( 2, 3, 4, 5, and 6) to characterize the hinge design features and performance. The intended application was to package several antennas, approximately twenty meters long, into a compact rectangular package. The deployment of this tube was mostly dependent on hinge torque, as there was little to no mass or rotational inertia at the tip relative to a system such as the Roccor PDS. Utilizing a similar tube and hinge architecture to package the Roccor PDS system would require an evaluation of hinge torque required to deploy a massive panel, as well as its ability to stabilize the panel mass once deployed. To assess the feasibility of the hinged boom for this application the deployed system was modeled in ABAQUS (Figure 4). Figure 4: Modal Evaluation of locked - out Panel Deployment System Results of this study showed that the first fundamental frequency mode was rotation about the boom, which cause the hinge tapes to shear relative to one another. Therefore, the two driving design requirements for the hinge in the Roccor PDS were deployment torque and shear stability. Optimization and validation of the hinge to provide enough deployment torque and torsional stiffness was the primary challenge for Roccor in the development of the panel deployment system.

Document

AMS_2020.pdf

386 Hinge Geometry and Design The geo metry of the hinge cutout in the foldable tube system was derived to both maximize torque output and reduce the high stress concentration areas that result from complex curvatures near the fold. By creating a narrow slit in the boom, the cross section effectively becomes two tape springs, which are commonly used in deployable structures for solar array applications. At the ends of the slit, however, the transition area must accommodate deformations to go from flattened tape springs to a full tube cross section. To avoid large strain areas due to the complexity of this transition area, a circular cutout was introduced. This cutout simply removes this difficult transition area. Figure 5: Foldable Hinge Geometry In the hinge cutout, there are several dimensions and material properties that effect the overall performance. These dimensions include the hinge length “L”, slot width “w”, end cutout diameter “d”, tube diameter “D”. The material properties of thickness, density, and modulus (viscoelastic and elastic contributions) must be accounted for to fully characterize hinge performance. Verification of Solid -State Hinge Design of the solid- state hinge for the ROC -PDS required verification testing and analysis of material allowable limits, hinge torque output, and torsional stiffness. For this, several testing methods had to be developed by Roccor to investigate the properties and effects of interest such as; material bending limits, hinge performance, and performance degradation due to long term holds. Material Testing for Thin laminate Bending As detailed in reference 7 the column bend test was developed to combine the attributes of both the platen (Ref. 8) and LD -FPBT (Ref. 9) into a simple test method intended to measure the moment -curvature relationship of a given HSC laminate. Testing is done by fixing a laminate sample into two identical “arms” that are then pinned into clevis mounts attached to a load frame. When the load frame compresses the fixture, the arms rotate about the pins requiring the sample to curve between the two arms. Using geometry, the moment -curvature relationships can be found using methods and equations detailed in Ref. 7. From this testing we can obtain the allowable bending curvatures, and relative strains, for a given laminate. For the ROC -PDS the laminate architecture selected contained a combination of Astroquartz and carbon fibers cured in a thermoset epoxy resin matri x. Testing was conducted on coupons cut at four orientations (0°, 45°, 90°, and 135°), and testing was conducted to obtain both the quasi -static load failure as well as the long- term hold failure curvatures ( Figure 7). Results from ultimate testing were p lotted on a polar chart for visualization of “allowable failure”. The average failure of each laminate orientation is utilized to create a polar curve fit “allowable failure” value to be used in analytical failure investigations of the stowed and deploying boom structure. Results from

Document

AMS_2020.pdf

387 relaxation testing were used to develop a custom viscoelastic relaxation curve using the Prony series method. Figure 6: Column Bend Test Fixture Geometry [ 10,11] Figure 7: Material Testing of composite hinge material. Figure 8: (Left) Curvature allowable values ; (Right) Laminate Prony series output

Document

AMS_2020.pdf

388 Where the curve fit to the equations is found using E quation 1 with allowable values s hown in Table 1. 𝐾𝐾𝜃𝜃𝑓𝑓=𝐾𝐾𝜃𝜃𝑓𝑓cos(𝜃𝜃)2+𝐾𝐾90𝑓𝑓cos(𝜃𝜃)2+�4𝐾𝐾45𝑓𝑓-2(𝐾𝐾0𝑓𝑓+𝐾𝐾90𝑓𝑓)�[sin(𝜃𝜃)cos (𝜃𝜃)]2 ( 1) Table 1 Instantaneous failure curvature results Orientation (degrees) Curvature (rad/m) (rad/m) (rad/m) 0 329.70* 0.73 N/A 90 328.82* 0.01 N/A -45 275.2 8.03 13.54 45 306.44 15.13 25.5 *Maximum curvature observed due to column bend test fixture limitations. No physical failure observed. Hinge Element Failure Evaluation The folded boom structure was evaluated for failure simply by taking the element level curvature values and converting them into the principal values of K 1, K2, and θ using Moore’s Law equations (equation 2- 4). For this effort, the principal curvature values were taken and re- plotted on the undeformed structure to more clearly illustrate the critical elements within the structure (Figure 9). 𝐾𝐾1= 𝜅𝜅1+𝜅𝜅2 2+��𝜅𝜅2−𝜅𝜅1 2�2 +�𝜅𝜅3 2�2 (2) 𝐾𝐾2= 𝜅𝜅1+𝜅𝜅2 2−��𝜅𝜅2−𝜅𝜅1 2�2 +�𝜅𝜅3 2�2 (3) 𝜃𝜃 =0.5∗atan�𝜅𝜅3 𝜅𝜅1−𝜅𝜅2 � (4) Figure 9: Converting local element curvatures to principal values using Moore's Law conversion equations. Hinge geometry for the ROC -PDS was initially selected to satisfy failure criteria . However , it was soon discovered that the driving design requirements would drive the hinge design. All hinge geometry trade study values were subsequently based upon the cri tical limits of the material testing.

Document

AMS_2020.pdf

389 Hinge Torsional Stiffness Trade Study Torsional stability of the hinge was shown to be a driving design factor of the ROC -PDS. To address what design parameters would affect the torsional stability of the hinge most a model was developed. The model that was used to evaluate the hinges torsiona l stiffness was conducted using ABAQUS explicit dynamic solver, where the hinge section was fixed in all directions at one end and a pure moment was applied at the tip of the boom about the tube axis ( Figure 10). Figure 10: Torsional Stiffness trade study model setup It was expected that the diameter of the end circle cutout controlled by the “d” variable shown in Figure 5 would be the primary driver of torsional stability due to it producing the smallest effective torsional cross section. This evaluation was conducted to evaluate the effects of both length and diametric cutout of the end cutouts. The curvature results were also compared to the material allowable values were then evaluated using the methods described earlier . Figure 11: Initial torsional stiffness evaluation based on hinge geometry and material allowable prediction Results of this investigation showed both that the risk of material failure due to torsional inputs was very low, and that the primary concern was geometric buckling behaviors. However, these buckling behaviors occur at least an order of magnitude beyond any operational load cases, so this risk was retired. The design output from this investigation was that minimizing the end cutout geometry and using a short slot cutout length result in the stiffest hinge geometry. Hinge Torque Output Trade Study Based on standard practice at Roccor with deployable structures, it is required that any deployment hinge provide enough torque to deploy the system, from any static position within the deployment , with a factor

Document

AMS_2020.pdf

390 of safety of two. The analysis model from the torsional stiffness investigation was utilized again for the torque output investigation, with the application of tip load changed and a small perturbation force added on the compressed hinge tape to promote a controlled buckling of the hinge ( Figure 12). Figure 12: Hinge torque trade study model setup To maximize the torque output of the hinge the “w” term from Figure 5 was minimized in order to maximize the potential energy of the hinge. Verification of this design consisted of first evaluating the material survivability by modeling the folded hinge with finite element s and using the evaluation methods described earlier. Figure 13: Initial torque output capabilities based on hinge geometry and material allowable prediction Results of th e analytical hinge studies were to select hinges with minimalized “d” and “L” terms for the hinge, however further investigations using a longer boom which allowed for more continuous deformation of the tube cross section resulted in the hinge getting deformed far passed its material allowable limits. The reason this was not seen in the initial investigation was due to the end conditions creating a rigid circle relatively close to the hinge cutout area, resulting in artificially positive margins for the area around the circular cutout. To address this issue the hinges were re- evaluated using a longer tube section ( Figure 14). This allowed for more natural curvature regions to form around the hinges.

Document

AMS_2020.pdf

391 Figure 14: Lengthened hinge geometry trade analysis model This effect was exacerbated by the ability for the full ROC -PDS system to seat itself into a system low energy state ( Figure 15), where the local curvatures around the hinge often resulted in negative margin. Analysis results did not conclude with a geometric result which closed for this case, however a design alteration to the overall system was devised to add rigidity to the loc al area and hold the hinge open. This design feature is known as a “power band” and is the topic of another study paper currently in progress at Roccor. Figure 15: (Left) High curvatures in full system analysis; (Right) hinge shift into low energy state. Hinge Torque and Long- term Storage Testing Full-scale hinged boom testing was used to validate the analysis results and determine the properties of the hinges where analytical techniques were unable to predict the behavior of this complex system. The primary test used was hinge torque testing under ambient temperatures and operational temperature extremes. This was coupled with long- term storage testing to evaluate material survivability and viscoelastic effect s on strain energy relaxation. A torque test fixture was developed and qualified to accurately measure the hinge output torque. This fixture is shown in Figure 16; the fixture uses torque cells to measure the output torque of each hinge as the boom is slowly allowed to deploy, resulting in a semi -static torque measurement. This fixture was qualified and validated by measuring the torque output of an easily modelled metallic leaf spring.

Document

AMS_2020.pdf

392 Figure 16: Structural testing of integrated solid- state HSC hinge The primary resistive torque in this design was due to a flat, flexible electrical harness inside the boom. The behavior of this harness depends intimately on the boundary conditions supplied by the boom, so testing of the harness alone proved impossible. Instead, the hinge torque was measured with and without the harness , and the harness torque was derived by comparing these two torque curves. These values were then used to derive and show deployment torque margin. A typical torque curve derived from this testing is shown below in Figure 17.

Document

AMS_2020.pdf

393 Figure 17 Typical hinge torque of proximal hinge (left) and distal hinge (right). Later tests at operational temperature extremes were enabled by constructing a thermal chamber around the torque test hardware. This testing was both used to directly verify the torque margin and validate the analytical predictions. The impacts of the viscoelasticity on the deployment kinematics were also a key risk, and due to a lack of standardized test methods and stress relaxation failure models these impacts are very difficult to analytically model. Long- term storage testing was therefore used to sim ulate the impact of storage before deployment. Full-length booms were stowed in the anticipated storage shape, as shown in Figure 18, and t he timetemper ature superposition relationship was used to reduce the storage time required from years to days . Figure 18 Long- term storage testing During and after storage the boom was inspected for damage. The corners of each hinge w ere at particular risk of stress rupture during long- term storage, so these areas were subjected to extra scrutiny. Detailed images of these areas are shown in Figure 19. After storage at elevated temperature the kick -off force, hinge torque, and final boom geometry was measured. These measured values were fed into the analytical model to better predict the deployment of the hinged boom after storage.

Document

AMS_2020.pdf

394 Figure 19 Inspection of the proximal hinge (left) and distal hinge (right). The results of this testing were then used to validate the analytical model , show compliance with requirements, and provide knock -down factors for post -storage deployments. In particular , the requirements for long- term storage survival and deployment torque margins were verified directly through test. Final Boom Laminate Architecture Features At the culmination of the development program, there were several design features resulting from detailed analysis and testing efforts done at Roccor. First, the hinges were strengthened by “power bands” in order to keep the slit from collapsing when the s ystem is required to be compressed beyond the material limit s of the hinge deformation. Next, the “razor backs” were added which are localized areas with adjusted thickness used to tune the required torque output of the hinge tapes to develop the required strain energy balance for the deployment. Last, the “lateral lines”12 were added in the boom, which are a localized laminate change which is utilized to tune the kickoff energy of the system. All these features were developed using the methods described in the previous section for material allowable calculations , as well as physical testing for validation of concepts and verification against flight qualification environments. Figure 20: ROC -PDS boom design features

Document

AMS_2020.pdf