page_content
stringlengths 25
9.66k
| type
stringclasses 1
value | source
stringclasses 12
values | page
int64 1
596
|
|---|---|---|---|
85 Figure 3. TO Isolator Strut Cross-Sectional View
During assembly, each spring is pulled back 4.123 cm (1.625 in) from its free length. The reactive force of this tensile load on the spring compresses the Center Rod onto the front Delrin® pad and preloads the assembly. During launch, when a large enough tension load is applied to overcome the preload, the center rod will unseat. After unseating as the spring extends further under still higher tensile launch loads, the relative motion of the Center Rod may travel 1.59 cm (0.625 in) until it hits another Delrin® pad stop on the Flanged Sleeve. The Delrin® pads were used so that titanium-on-titanium contact would not occur during impacts, preventing galling of the contacting parts. The expanded diameter head on the front end of the Center Rod limits its motion to this 1.59 cm (0.625 in) travel space. The Ares first stage forcing function has been computed to apply tensile loads resulting in head travel within this 1.59 cm (0.625 in) range, effectively isolating the pallet from the rest of the crew module in the X direction. All of the load bearing machined parts and the machined spring are made out of titanium to reduce the weight of the assemblies. Each of the machined springs were machined from a single piece of titanium and have identical bolted interfaces on both ends. Due to the method of construction and the bolted interfaces, the springs can react both compressive and tensile loads. The 105 kN/m (600 lbf/in) Isolator spring assemblies each have a mass of 5.08 kg (11.2 lbm) and the 210.2 kN/m (1200 lbf/in) Isolator spring assemblies each have a mass of 6.21 kg (13.7 lbm). The total mass for all four Isolator spring assemblies is 22.6 kg (49.8 lbm) which does not include the existing CM portion of each strut. Damping To provide for the low-level damping required to maintain stability in a LA scenario, the motion of the Isolator requires a damping component. Due to the late addition into the design, a passive, friction-based damping system was devised. Two Delrin® components clamp on the Center Rod; fasteners squeeze the components together, providing a friction force against the Center Rod that can be controlled based on the fastener torque. They are designed to provide between 111 – 156 N (25 – 35 lbf) of frictional force. Lockout Mechanism In order to satisfy the locking requirement for landing loads, a feature was added to prevent any stroking of the Isolators. For ease of prototype testing and assembly, quick release (insertion) pins were inserted behind the head of the Center Rod preventing it from moving and creating a rigid load path that bypasses the machined spring. For flight, a quick-reacting lock out mechanism would be required during landing. A concept for a Non-Explosive Actuator (NEA) driven blade was added to the design. The concept uses an NEA to hold a spring-loaded Plunger that is threaded into the NEA nut; when actuated by a 4-amp electrical signal, the NEA releases the Plunger which wedges behind the Center Rod, securing the Rod in place and creating | Document | AMS_2012.pdf | 99 |
86 a single shear load path that bypasses the spring creating a strut stiffness equal to that of the CM baseline strut. A second blade 180° from the first would be needed for redundancy. The concept used for testing incorporated an NEA mockup on one of the four Isolators to characterize its static performance. Refer to Figures 4 and 5 for images of the Isolator and NEA Lockout Mechanism.
Figure 4. TO Isolator Strut with NEA Lockout Mechanism
Figure 5. NEA Lockout Mechanism Cross-Sectional View Test and Evaluation Program
The test and evaluation program aimed to characterize the effectiveness of the Isolation springs in attenuating the launch dynamic accelerations and loads imparted to the crew pallet. In addition, the test program aimed to understand the effect the Isolators had on the wire bender stroke during landing impact and better understand the correlation of model predictions to test results. The test program included: Isolator Proof Testing, Isolator Stiffness and Damping Characteristics, Strut Impact Testing, System Drop Testing, Modal Testing of the launch configuration, and Vibration Testing to characterize performance. NEA Lockout Mechanism (x2) Plunger Plunger drops behind Center Rod flange, locking Center Rod and directing load path through internal components Internal nut separates when NEA actuated, releasing Plunger NEA Center Rod Isolation Spring Coulomb Damper | Document | AMS_2012.pdf | 100 |
87 TO Isolator Proof Testing / Stiffness and Damping Characteristics Each Isolator assembly was tested in tension (locked and unlocked) and in compression to prove strength margins and verify spring stiffnesses. Maximum loading expected on the strut occurs during drop testing and is 4.448 kN (10,000 lbf). A test factor of 1.25 was included in proof tests leading to testing in compression and tension at 5.560 kN (12,500 lbf). All proof tests were performed successfully with no signs of failure or yielding. Spring stiffnesses and friction forces were found to be in acceptable ranges. Strut Sub-System Impact Testing Impact testing aimed to characterize TO vibration isolation springs in a dynamic environment and assess their effect on the behavior of the wire bender landing attenuation struts. In addition, an accurate LS-DYNA® model of the vibration isolation spring and attenuation strut was desired so that a reliable analytical model is available for subsequent landing simulations performed with the Orion vehicle under simulated landing conditions.
The tests were conducted using the 711th Human Performance Wing’s Vertical Deceleration Tower (VDT) at Wright Patterson Air Force Base in Dayton, Ohio. The VDT consists of an 18 m (60 ft) vertical steel tower that allows a carriage to enter a free-fall state (guided by rails) from a pre-determined drop height. The plunger mounted on the rear of the carriage is guided into the hydraulic deceleration device (cylinder filled with water located at the base and between the vertical rails), producing an impact deceleration pulse, see Figure 6.
Figure 6. Attenuation Struts in VDT Setup
The testing showed that the transmitted acceleration peaks are similar regardless of the presence of the isolation spring. Testing and simulations for the strut with both the TO vibration isolation spring and the wire bender incorporated into the strut showed that the isolation spring has the effect of reducing the Brinkley injury criteria at the expense of increasing the strut stroke.
System Drop Testing Orion CIAS impact tests were performed at the NASA Langley Landing and Impact Dynamic Research Facility to provide a demonstration of strut system performance, evaluate the performance differences VDT Carriage with Strut Wire Bender Isolator | Document | AMS_2012.pdf | 101 |
88 between the locked and unlocked condition of the TO Isolators during the landing impacts and assess LSDYNA® modeling techniques and predictive capability. The test configuration had a crew pallet-mockup (green, Figure 7) suspended from a cage (yellow, Figure 7) via the eight struts simulating the Orion CM interior. The cage was housed in a ring base/support structure to provide adjustment for various landing parachute pitch angles. Vertical drop tests were conducted by releasing the test fixture from a crane hook at a height calculated to produce the desired impact velocity. Tapered stacks of paper honeycomb at the four corners of the test fixture base were used to produce impact pulses approximating Orion water landings (see Figure 7). Measurements recorded during the tests include TO Isolator displacements, forces and accelerations as well as pallet and cage accelerations. High-speed cameras and photogrammetry were used to verify impact conditions and observe TO behavior. Eleven CIAS system drop tests were performed, which successfully demonstrated the performance of the system of struts and provided data for evaluation of the effect of the locked/unlocked condition. The accuracy of the LS-DYNA® model was also assessed. The tests featured impact velocities ranging from 3.05 - 10.7 m/s (10 - 35 ft/s) with the crew pallet locked at a 28 o pitch angle. The findings from this test and simulation effort are as follows: 1. Strut force and pallet acceleration time histories can be predicted via LS-DYNA® simulations with a high degree of accuracy and are relatively insensitive to expected variations in strut parameters such as strut load limit levels, initial stiffness, and dead zones (initial slack). The load limit in the struts determines the peak acceleration of the pallet. The expected range of the strut force limits will result in a relatively minor variation in the strut forces and pallet accelerations. 2. The strut stroke is the most important parameter to consider for evaluation of the system response. It is also the most difficult output to predict, due to its high sensitivity to most input variables. Comparisons of test data with the LS-DYNA® simulation results for tests 3 through 11 had an average prediction error in the strut displacements of ±0.66 cm (0.26 in). The largest observed strut displacement error between a test and simulation was 3.8 cm (1.5 in). The overall average error was 20%. Accurate prediction of the strut strokes requires a high level of fidelity in the modeling of the structure to capture the flexural response of the crew pallet and the structure supporting the outboard ends of the struts, as well as very accurate modeling of the energy-absorbing wire bender strut force versus displacement curve. 3. Depending on the ratio of the load limit magnitude of the wire bender struts and the stiffness of the TO Isolator, there can be an amplification of the wire bender strut strokes for the unlocked condition. The testing revealed that there are combinations of wire bender struts and Isolator struts where the unlocked condition of the Isolator struts does not result in amplification of the wire bender strokes, and other combinations that can amplify the wire bender strokes by a factor of 2.5 to 3.0. These results confirmed the LS-DYNA® predictions that the Isolator needed to be in a locked configuration during landing. 4. The LS-DYNA® model is accurate enough to be used as an effective design tool for further CIAS studies. The design uncertainty on the pallet acceleration environment will reflect the expected variation in the strut yield force and is expected to be limited to 10% provided that the struts do not exceed their stroke limits. For the strut strokes, a design margin of 20% should be used. | Document | AMS_2012.pdf | 102 |
89 Figure 7. Attenuation Strut in Assembly (Left); Drop-Test Fixture (Right)
System Modal Testing As part of the pre-test planning for the base-drive vibration testing of the CIAS isolation system, a series of modal tests were performed at NASA Langley to assess the pre-test finite element model (FEM) predictions. The objectives of the modal test were: (1) to investigate potential fixture modes in the frequency range of the vibration test (0 - 20 Hz); and (2) to verify FEM predicted modes for the isolated crew pallet. Two Isolator test configurations were evaluated: (1) locked and (2) unlocked. In the locked configuration, the isolated struts perform like rigid elements. For the unlocked configuration, the Isolators were shimmed to position with the Isolators in their active stroke range. Accelerometers measured the Frequency Response Functions (FRFs), calculated as the ratio of the acceleration response to the input force. Modal parameters (natural frequencies, damping factors, and mode shapes) were then estimated from the FRFs. Base-drive data acquired during the vibration testing was also used to obtain modal estimates with and without the friction dampers, see Figure 8.
Figure 8. LS-DYNA® Model (Left); System-Level CM Mockup (Right) Due to low-level inputs, at certain times it was difficult to overcome the friction in the damper, which created difficulties in test execution and data collection. Eventually the friction dampers were removed, allowing for more ideal performance of the Isolators; as a result all three pallet modes were identified. Crew Pallet Mass Mockup Wire Bender & Isolator Strut Crushable Honeycomb Support Structure CEV CM Structure Mockup | Document | AMS_2012.pdf | 103 |
90 There is good agreement between the measured frequency estimates and FEM predictions with the largest frequency difference at 8% for the first mode. Modal results are presented in Table 1. Table 1. Modal Estimates for the CIAS Pallet Modes Mode Predicted (Hz) Modal with friction dampers Base-Drive with friction dampers Base-Drive without friction dampers Freq (Hz) Damp (%) Freq (Hz) Damp (%) Freq (Hz) Damp (%) Rocking-Y:+Z struts 3.77 4.1 8.8 - - 4.1 3.2 Twist about X 4.37 - - 4.5 7.9 4.7 4.1 Rocking-Y: -Z struts 5.35 5.3 6.0 5.3 7.1 5.4 3.5 System Vibration The objective of the system-level vibration test was to demonstrate the effectiveness of the pallet-to-strut isolation springs for reducing crew loads during a TO event occurring at approximately 12 Hz. The Orion wire bender Vibration Testing Unit (VTU) was subjected to dwell tests at frequencies of 10 and 12 Hz with varying amplitudes to represent a range of possible launch oscillation loads. Testing occurred at the Naval Surface Warfare Center (NSWC) Dahlgren Division Vibration Test Facility. See Figure 9 for a picture of the test setup. The test consisted of three configurations: Test Sequence 1: rigid, Isolators locked out Test Sequence 2: Isolators released, friction dampers installed Test Sequence 3: Isolators released, friction dampers removed
Each configuration was subjected to three sine sweeps at varying input levels to test linearity. After the sine sweeps, each configuration underwent dwell testing at the TO frequencies with three different input levels: 0.2, 0.35, 0.5g. Conducting the signature sine sweeps verified the linearity of the system for Test Sequences 1 and 2. Because dampers were removed for Test Sequence 3, acceleration limits were reached as the pallet isolation frequency coupled with the input. As a result, only one successful sine sweep at 0.025g was completed over the range 2 – 8 Hz for Test Sequence 3. The test was modified to only sweep across 8 – 20Hz for levels higher than 0.025g (0.075 and 0.1g) which verified the linearity of the system over the range 8 – 20 Hz. However, the linearity of Test Sequence 3 over the range 2-8 Hz cannot be verified. Any slight deviation from linearity in Test Sequence 1, Test Sequence 2, and Test Sequence 3 (8-20 Hz only) can be attributed to the noise introduced by the hydraulic shakers. Figures 10 and 11 display the reduction in accelerations attributed to the Isolators at the 10 Hz dwell and 12 Hz dwell. Tables 2 and 3 display the reduction in acceleration for the 10 Hz and 12 Hz dwells (note that a negative reduction indicates an increase in acceleration). | Document | AMS_2012.pdf | 104 |
91 Figure 9. Orion VTU Mounted on Hydraulic Shakers in the NSWC Vibration Facility
Figure 10. Comparison of Pallet 10 Hz Dwell Test Results for Each Test Configuration
00.10.20.30.40.50.60.70.8 0.05 0.2 0.35 0.5 0.65Acceleration Output (g) Acceleration Input (g)Accel 1 - TS1 (Rigid) Accel 2 - TS1 (Rigid) Accel 3 - TS1 (Rigid) Accel 4 - TS1 (Rigid) Accel 1 - TS2 (Isolated w/ Damper) Accel 2 - TS2 (Isolatedw/ Damper) Accel 3 - TS2 (Isolatedw/ Damper) Accel 4 - TS2 (Isolatedw/ Damper)10 Hz Dwell Tests Pallet Comparison Plot | Document | AMS_2012.pdf | 105 |
92 Figure 11. Comparison of Pallet 12 Hz Dwell Test Results for Each Test Configuration
Table 2. Average Pallet Accelerations for 10 Hz Dwell Tests G Input Test Sequence Input as Measured (g) Average Pallet Response (g) Average Pallet Transmissibility Average Pallet % Reduction Compared to Input* Average Pallet % Reduction Compared to TS1 (Rigid) 0.2 TS1 0.203 0.258 1.27 -27.09 N/A TS2 0.2 0.092 0.46 54.00 64.34 TS3 0.198 0.06 0.30 69.70 76.74 0.35 TS1 0.348 0.443 1.27 -27.30 N/A TS2 0.347 0.133 0.38 61.67 69.98 TS3 0.355 0.112 0.32 68.45 74.72 0.5 TS1 0.501 0.642 1.28 -28.14 N/A TS2 0.506 0.18 0.36 64.43 71.96 TS3 0.497 0.154 0.31 69.01 76.01 *Note: Negative reduction values indicate increase in input. 00.10.20.30.40.50.60.70.80.9 0.05 0.2 0.35 0.5 0.65Acceleration Output (g) Acceleration Input (g)Accel 1 - TS1 (Rigid) Accel 2 - TS1 (Rigid) Accel 3 - TS1 (Rigid) Accel 4 - TS1 (Rigid) Accel 1 - TS2 (Isolated w/ Damper) Accel 2 - TS2 (Isolatedw/ Damper) Accel 3 - TS2 (Isolatedw/ Damper) Accel 4 - TS2 (Isolatedw/ Damper)12 Hz Dwell Tests Pallet | Document | AMS_2012.pdf | 106 |
93 Table 3. Average Pallet Accelerations for 12 Hz Dwell Tests G input Test Sequence Input as Measured (g) Average Pallet Response (g) Average Pallet Transmissibility Average Pallet % Reduction Compared to Input* Average Pallet % Reduction Compared to TS1 (Rigid) 0.2 TS1 0.198 0.281 1.42 -41.92 N/A TS2 0.197 0.106 0.54 46.19 62.28 TS3 0.199 0.042 0.21 78.89 85.05 0.35 TS1 0.347 0.52 1.50 -49.86 N/A TS2 0.35 0.122 0.35 65.14 76.54 TS3 0.344 0.073 0.21 78.78 85.96 0.5 TS1 0.507 0.776 1.53 -53.06 N/A TS2 0.5 0.147 0.29 70.60 81.06 TS3 0.492 0.105 0.21 78.66 86.47 *Note: Negative reduction values indicate increase in input.
During Test Sequence 3, the average reduction in acceleration or loading remains relatively constant over all the input levels tests. Test Sequence 2 on the other hand seems to experience a greater percent reduction as the input level is increased. After the completion of Test Sequence 2, the Isolators were removed and the friction damper force was tested. The friction dampers for Isolators 7 and 8 were within the 111 – 156 N (25 – 35 lbf) target range while struts 5 and 6 had static/kinetic friction values of 267 N (60 lbf) and 67 N (15 lbf), respectively. During low level dwell and sine sweeps of Test Sequence 2, strut 5 was not fully released which resulted in a shift in the system dynamics. The test demonstrated that the TO Isolators do an effective job at mitigating loads due to a TO event, reducing pallet accelerations to 20 - 40% of that of the input. The Delrin® friction dampers however, are overly sensitive and altered the dynamics of the system making correlation of pre-test analyses with the dampers difficult. This prototype damping system should be replaced with a more reproducible, controllable damper for a flight system.
Lessons Learned
Passive Friction Damping The Isolator design solution utilized a passive, friction-based damping scheme. The inclusion of damping into the Isolator was not introduced until after fabrication of the Isolator parts, which partially drove the solution. Friction, by nature, is difficult to control and get repeatable results. The end result of this is that the friction force applied to the Center Rod, and therefore the damping coefficient of the Isolator, changes slightly from test to test. In addition, the Center Rod was turned down on a lathe, leaving a concentricity and roundness tolerance greater than desired. This caused variations in friction clamping force throughout Center Rod travel, which was observed in Isolator friction testing. A more appropriate finish on the Center Rod for a passive friction damping system would be to centerless grind the shaft, which would result in less surface variations and more consistent friction forces. Structural Adhesive under Impact Loads At both the forward and aft hard stop of Center Rod travel, a Delrin® piece would contact the Center Rod to prevent like-material contact. Hysol® EA 9394 structural adhesive epoxy was used to adhere the Delrin® to the substrate metal. This was selected due to the fact that it would require minimal modification to the components, ease of assembly, low-profile installation, and common usage in space industry. The | Document | AMS_2012.pdf | 107 |
94 forward Delrin® pad was bonded to the inner counter-bore on the Front Hub. The aft Delrin® pad was bonded to the rear of Center Rod head. The Delrin® pads did not dislodge or displace during proof testing. In the dynamic tests (such as Strut Impact Testing, System Drop Testing, and System Vibration) it was observed after certain test runs that the Delrin® would no longer be bonded to its substrate metal. The high impact loads would break the bond causing the Delrin® component to be loose within the cavity. Therefore it has been observed that structural adhesives are likely not appropriate for dynamic impact loading. If it were to be re-designed, the Delrin® pads would be hard mounted with small flat head screws slightly recessed below the impact surface. Displacement Data Capture in Dynamic Environment As a critical parameter used for performance and model correlation, the testing program required displacement measurements to be captured during an impact. String-potentiometers were installed in parallel to the struts to record the motion of the impact attenuation wire benders with very good performance. But when installed to measure isolator displacements during single strut/isolator impact tests, the string pots demonstrated greater error band than expected; often giving measurement readings that were much higher than theoretical maximums. From examining high speed camera video, the string pots appeared to have insufficient response during impact onset resulting in over deployment during initial extension and out-of-axis string movement during retraction. The exact cause of this is indeterminate; suspected causes are inertial affects within the device making it inadequate for isolator dynamics during impact or lack of stiffness in the string-pot mounting brackets. Linear Variable Differential Transformers (LVDTs) using rigid plunger shafts were implemented for system drop and vibration tests and were found to perform well with reliable isolation deflection data.
Data Acquisition System Updating Data from the Orion vibration test was captured using two data acquisition systems (DASs) with different capabilities. The Modal DAS was used to acquire the acceleration measurements on the test article. This DAS was specifically designed to capture and process data for near real-time viewing during the test. For the strut load cell and LVDT data acquisition an EME Corporation Model 3200L DAS was utilized. The EME DAS was designed to capture time data only and did not have any processing capability. The EME Corporation Model 3200L DAS could only be set to acquire data based on a fixed time interval. If a test ran longer than expected, the EME DAS time may expire requiring a re-test. As a result, the EME DAS was set to have a time approximately 5 minutes longer than the estimated testing time. The problem is, once the EME DAS starts, it cannot be stopped until the fixed time expires. This led to gaps between tests and acquiring data past the test’s completion. It is recommended that an updated DAS be used for future tests, which do not rely on user defined data acquisition time frame, to save time between tests and to eliminate the acquisition of data past the test’s completion.
Conclusions
The TO study confirmed the optimal crew isolation frequency of 4.5 Hz and testing established the system performance and damping mechanism value. From a load mitigation perspective, it was found that the pallet isolation approach was very appealing. Results indicated that the isolation system provided a reduction of dynamic load to about 20% - 40% of the input and that Brinkley levels were met at a mass penalty of less than 5.9 kg (13 lbm) per strut.
The results of this test program illustrate the feasibility and benefits of implementing a pallet isolation system for the Orion CEV CM. The design and test data included herein are directly applicable to the Orion vehicle, but could be adapted to other designs with similar dynamic load issues. Isolation for load reduction is a flight proven technology utilized on several robotic spacecraft in addition to Space Transportation System payloads. Its low mass penalty is relatively insignificant when compared to the hardware benefits and the potential mass increases if this option is not exercised. | Document | AMS_2012.pdf | 108 |
95 Acknowledgements
The authors would like to thank Chuck Lawrence, Greg Walsh, Steve Mark, Steve Hendricks, Justin Templeton, and Don Jarosz for their work and support on this project and acknowledge the invaluable contributions from the test teams at Wright Patterson AFB Impact Test facility, Dahlgren Naval Facility, as well as NASA Langley's Landing and Impact Research Facility and Vibration/Modal Laboratory. Their support during this project was essential to its success.
References
1. Lawrence, Charles et al. NASA/TM-2008 215198. The Use of a Vehicle Acceleration Exposure Limit Model and a Finite Element Crash Test Dummy Model to Evaluate the Risk of Injuries During Orion Crew Module Landings. April 2008. | Document | AMS_2012.pdf | 109 |
97 The Damper Spring Unit of the Sentinel 1 Solar Array
Frans Doejaaren* and Marcel Ellenbroek*
Abstract
The Damper Spring Unit (DSU, see Figure 1) has been designed to provide the damping required to control the deployment speed of the spring driven solar array deployment in an ARA Mk3 or FRED based Solar Array in situations where the standard application of a damper at the root-hinge is not feasible. The unit consists of four major parts: a main bracket, an eddy current damper, a spring unit, an actuation pulley which is coupled via Kevlar cables to a synchro-pulley of a hinge. The damper slows down the deployment speed and prevents deployment shocks at deployment completion. The spring unit includes 4 springs which overcome the resistances of the damper and the specific DSU control cable loop. This means it can be added to any spring driven deployment system without major modifications of that system. Engineering models of the Sentinel 1 solar array wing have been built to identify the deployment behavior, and to help to determine the optimal pulley ratios of the solar array and to finalize the DSU design. During the functional tests, the behavior proved to be very sensitive for the alignment of the DSU. This was therefore monitored carefully during the qualification program, especially prior to the TV cold testing. During TV “Cold” testing the measured retarding torque exceeded the max. required value: 284 N-mm versus the required 247 N-mm. Although this requirement was not met, the torque balance analysis shows that the 284 N-mm can be accepted, because the spring unit can provide 1.5 times more torque than required. Some functional tests of the DSU have been performed without the eddy current damper attached. It provided input data for the ADAMS solar array wing model. Simulation of the Sentinel-1 deployment (including DSU) in ADAMS allowed the actual wing deployment tests to be limited in both complexity and number of tests. The DSU for the Sentinel-1 solar array was successfully qualified and the flight models are in production.
Figure 1: Damper Spring Unit.
* Dutch Space BV, Leiden, The Netherlands
Proceedings of the 41st Aerospace Mechanisms Symposium, Jet Propulsion Laboratory, May 16-18, 2012 | Document | AMS_2012.pdf | 111 |
98 Introduction
The Damper Spring Unit (DSU) is part of the deployment system of the Sentinel 1 Solar Array (SA), an ARA Mk3 product family solar array from Dutch Space. Sentinel-1 is an imaging radar satellite aimed at providing continuous all-weather, day-and-night imagery for monitoring sea ice zones and the arctic environment, surveillance of marine environment and monitoring land surface motion risks, mapping of land surfaces: forest, water and soil, and provide mapping in support of humanitarian aid in crisis situations. The function of the deployment system is to deploy the Sentinel 1 SA wing from the folded panel stack (stowed configuration) mounted on the S/C side wall to the deployed configuration (see Figure 2). The DSU is designed to provide the damping required to control the deployment speed of the spring driven solar array in situations where the standard application of a damper at the root-hinge is not feasible. The large number of power harness cables of the Sentinel 1 SA prevents the location of the damper at the root hinge. This paper gives a description of the damper spring unit and its functioning in the Sentinel 1 deployment system . The qualification program of the DSU is discussed, the torque balance analysis and deployment analysis in ADAMS, which is required to qualify the use of the DSU in the Sentinel 1 deployment system.
Figure 2: Sentinel 1 solar array wing (stowed and deployed configuration).
Sentinel 1
The DSU will be first applied in the Sentinel 1 solar array wing. The wing consists of 6 panels; 5 solar panels and a yoke panel. It is fixed to the spacecraft via 4 hold down stacks, each containing 2 Thermal Knives to cut the hold down cable inside the hold down stack (see Figure 2 and Figure 3). The deployment system of the Sentinel 1 SA consists of: Twelve (12) panel hinges per wing (i.e. two hinges per hinge line). Each hinge has a pretensioned spring to deliver the hinge line torque for deployment and a locking device. A deployment synchronization system which consists of pulleys, guide-blocks and cables to control the deployment. A system of Kevlar synchronization cables are connected via pulleys to the panel hinge assemblies. These cables synchronize the panel motion between all panels. The damper-spring-unit, mounted on the yoke panel front side (at sun side in deployed condition and between yoke panel and spacecraft side wall in stowed condition) and coupled via Kevlar cables to the synchro-pulley of a dedicated DSU hinge at hingeline 2 (see Figure 4). Kick-off-spring-unit attached on the yoke panel and connected with the root hinge synchronization pulley to support a compact first stage 90° deployment. | Document | AMS_2012.pdf | 112 |
99
Figure 3: Sentinel 1 stowed solar array wing (side view).
The Sentinel 1 solar array deviates slightly from standard ARA Mk3 solar array designs mainly in respect with the deployment movement from stowed to the deployed configuration. The root hinge must deploy over 180 degrees, but more important, the trajectories of the C.O.G.’s of all panels have to be confined to a limited area which requires a two stage deployment movement of the panels which is shown in Figure 5. The synchronization system controls these 2 phases in a continuous way avoiding interference with S/C structure. The Sentinel 1 deployment design is a passive system as the traditional deployment system with the same design of hinges, motoring springs, synchro-cables and eddy-current deployment damper. The root hinges pulleys are, however, specifically designed eccentric and the routing of the cables is adapted. The deployment of the panels starts immediately after the last hold down cable has been cut. The deployment motion is driven by the kick-off spring on the yoke panel and the deployment springs mounted at the root hinge and at the panel hinges. The eccentric roll-on / roll-off pulleys on the root hinges realize a gradually increasing synchro-cable movement. The cable payout up to 90 degrees hinge line rotation is small, thus the panels will stay approximately stacked up to about 90º opening of the root hinge line (phase 1). In phase 2 the root hinge opens from about 90 to 180 degrees, while all other hinges open in a synchronized manner up to deployment completion (from about 5 to 180 degrees). This sequence ensures that the CoG travel and MoI shift – and therefore the disturbances on the spacecraft’s attitude - are minimized. At the end of the deployment each individual hinge will latch to assure a stiff deployed configuration.
Figure 4: The DSU position at the Sentinel 1 solar array. | Document | AMS_2012.pdf | 113 |
100 Sentinel 1 Engineering Model
Using engineering models helped us to identify the behavior of the two stage deployment system and how the DSU functions within this system (see Figure 5). Due to the severe thermal environment, some flexibility of the synchronization system is required. This flexibility also introduces unwanted non-synchrony. Design modifications are introduced to minimize these effects. As the DSU is not directly integrated in the synchronization system, its pulley can be increased or reduced in diameter. This change can introduce a more favorable ratio between the various pulleys and thus limit the non-synchrony effects of the flexibility. Unfortunately these effects cannot be eliminated completely, which means it has to be accounted for in the deployment analysis of the system. After the final pulley ratio was identified, the DSU design could be finalized.
Figure 5: The deployment of the Sentinel 1 engineering model. | Document | AMS_2012.pdf | 114 |
101 Damper Spring Unit description
The Damper Spring Unit consists of 4 major parts (see Figure 6): The main bracket, which supports all other parts of the DSU and allows fixation to the Solar Array. The main bracket interface with the solar array is dedicated per project. The Sentinel 1 DSU is fixed with a M4 into a blind insert and a M8 bolt which is interfaces with an adapted cup-cone bracket, see Figure 7. The spring unit, which includes 4 standard ARA Mk3 root hinge actuation springs, to counter the retarding torque of the specific DSU control cable loop and the eddy current damper start-up torque with sufficient factor of safety. The actuation pulley, which is coupled via Kevlar cables to a synchro-pulley of a hinge of the solar array (see Figure 9). The damper is an eddy current damper supplied by RUAG Space AG, Zurich, Switzerland. The damper consists of three basic modules: o A high efficiency damper unit which is a self-contained module with its own shaft supported by its own set of bearings. The damping is generated with a high purity copper disc rotating within a highly concentrated magnetic field. The field is provided by 12 pairs of samarium-cobalt magnets. The damping rate can be varied. The different damping rate levels depend on the relative orientation of the magnet pairs which can easily be set by rotating the ECD end cover. o An intermediate gear-head is employed based on a standard planetary unit. This unit provides the first stage of torque amplification. o The input stage is the second torque amplification stage and provides a further amplification ratio. This stage is designed to accommodate high torque levels (up to 100Nm). The damper is currently qualified to provide a damping rate between 500 and 1700 N-m-s/rad at operational temperatures of -55 to +100°C. A dedicated attachment plate is required to attach the damper to the main bracket.
Figure 6: Damper Spring Unit parts. | Document | AMS_2012.pdf | 115 |
102
Figure 7: Damper Spring Unit interface.
Due to the heavy cable loads typical for a DSU application a new cable end fitting has been designed and qualified. The Kevlar cable is a standard ARAMk3 / FRED cable type of which the fitting is changed from a single knot in a cylindrical bush into a spike in a conical receptacle (see Figure 8). Both DSU cables contain a preload spring. The design of the preload spring depends on the routing of the synchro-cable. For the Sentinel 1 DSU cable route a preload spring with a preload of 5N has been adopted. This value is an important aspect in the torque balance of the unit as a whole and should be analyzed carefully. The torque balance analysis is described later in this paper. Figure 9 shows the interfaces of the DSU cables.
Figure 8: DSU control cable design. Figure 9: DSU control cable interfaces. | Document | AMS_2012.pdf | 116 |
103 The Center shaft design is less straightforward as one would expect due to the thermal environment requirements and the variation in thermal coefficient of the damper stainless steel input shaft and the aluminum 7075 Center shaft. The damper has a steel input shaft unlike the Center shaft, the main bracket and the cable pulley, which are of aluminum. An additional stainless steel shaft, the damper engagement shaft is introduced to interface with the damper input shaft and the Center shaft. The damper engagement shaft is coupled with the Center shaft by the lock bolt shown in Figure 10.
Figure 10: Damper Spring Unit shaft design. Qualification program The qualification program consisted of various functional tests at ambient and at the extreme temperatures (-115 ˚C / +115 ˚C), static load tests, random- and sine-vibration tests and a life test. As the damper has been qualified by RUAG, not all tests are performed with the damper attached to the main bracket. This allows a full characterization of the DSU, which is required to create the ADAMS model. Figure 11 gives the test sequence of the DSU, which indicates when the damper was not used.
Figure 11: DSU qualification test sequence. | Document | AMS_2012.pdf | 117 |
104
Figure 12: Functional test of the DSU qualification model (ambient).
Random vibration tests have been done and the test requirements of Table 1 and 2 have been met.
Table 1: Random Vibration Test requirements (all 3 axes)
Frequency Range [Hz] PSD Qualification 10 – 80 + 6 dB/oct until 0.6 g²/Hz 80 – 400 0.6 g²/Hz 400 – 2000 - 6 dB/oct Overall random vibration level [G RMS ] 20 Duration [seconds] 120
Table 2: Sine Vibration Test requirements (all 3 axes)
Frequency Range [Hz] g-level 5– 22 Maximum shaker amplitude 22 – 100 30
During the functional tests, the behavior proved to be very sensitive for the alignment of the DSU. This was therefore monitored carefully during the qualification program, especially prior to the TV cold testing. During TV “Cold” testing the retarding torque exceeded the max. required value: 284 N-mm versus the required 247 N-mm. This requirement had been set conservative and fortunately the torque balance analysis shows that the 284 N-mm is acceptable. Because this Non Conformance has been resolved and all other pass/fail criteria are met, like the ultimate load of cable A of 3360 N, the DSU qualification test program has been successful and the DSU design is qualified for use within the specified environment (temperatures, forces and number of on ground cycles). | Document | AMS_2012.pdf | 118 |
105 Torque Balance Analysis
The torque balance analysis presented here shows the torque balance for the critical situation of the deployed position during cold deployment. In the cold condition the friction coefficients and the damper retarding torque are at their highest, and the deployed position has the lowest available DSU drive spring torque. Figure 13 shows the torque balance calculation flow and Figure 14 shows which cable forces and deflection angles are taken into account. The torque balance budget for the DSU and its control cables, given in Table 3, shows that the DSU motoring torque is a factor 3.3 higher than the worst case retarding torque. A motoring factor (torque margin) >3 is required, so this requirement is met. The DSU spring unit can provide up to 7400 N-m, consequently a higher motoring factor can be achieved if so desired.
Figure 13: Torque balance calculation flow.
Figure 14: DSU cable forces and deflection angle definitions. | Document | AMS_2012.pdf | 119 |
106 Table 3: The DSU system torque balance budget.
Sentinel 1 Deployment Simulation in Adams
The deployment of the Sentinel-1 solar array is studied with the commercial multi-body package ADAMS which is part of the mechanical engineering software of MSC. In ADAMS, a model of a solar array wing is characterized by Parts, Joints and Loads. The ADAMS Sentinel 1 model is shown in Figure 15. The parts in the model description are: Panels: the yoke panel and the solar panels. Synchronization system: pulleys, guide-blocks and begin and end parts of the synchronization cable. The damper unit The constraints in a solar array wing model define: revolute joints (or hinges) between the panels Constraints relating two markers that together define a marker for the (virtual) synchro-cable (The contributing load is implemented and a graphical element. No actual body is representing the cable) The loads in the model are: Deployment torques, consisting of o Torques to deploy the stack of panels o Kick-off spring which pushes the complete stack during the first 60 degrees of the root hinge Non-linear loads to prevent that the hinges between the panels rotate beyond the structural boundaries (The panels are not able to cross each other) Linear loads activated when the hinge locks (=at the end of the hinge deployment) The damper load applied by the damper The loads in the synchronization cables COLD CASE DEPLOYMENT Torque Safety Factored factor torque [Nmm] [-] [Nmm] RETARDING TORQUE DSU DEPLOYED damper start torque -1240 3 -3720 Resistiv e torque of the DSU shaft + pulleys -284 3 -852 TOTAL retarding -1524 3 -4572 TOTAL motoring torque DSU springs 5100 1 5100 Motorization factor 3.3 | Document | AMS_2012.pdf | 120 |
107
Figure 15: The Sentinel 1 SA wing representation in ADAMS.
The functional components are modeled in ADAMS as User Defined Entities (UDE’s). Since the functional components are fully parameterized the UDE’s can be re-used in a very efficient way. The actuation springs are straightforwardly represented by moments in the hinges. The characteristic moment – angle behavior is introduced as a prescribed curve. Unfortunately, in ADAMS cables and pulleys are not standardly available and therefore a toolbox developed by Sayfield International has been used. With this toolbox the synchronization system has also been modeled as a UDE with help of sub-models of the pulleys, slip-rings and cables from the toolbox. In these sub-models, the cables are not physically introduced but as forces. The spring damper unit is also modeled with the cable toolbox. Figure 16 shows the complete DSU system: the damper disk, where the damping load is applied, the outer disk, the cables and the pulley. Only the spring unit of the DSU is not shown. The spring is modeled by a moment as a function of the rotation angle of the outer disk.
Figure 16: The DSU representation in ADAMS. | Document | AMS_2012.pdf | 121 |
108 The two stage deployment of Sentinel 1 is shown in Figure 17. Such a trajectory is achieved by placing the pulleys off centered and using different sizes for the pulley radii. This adapted design changes the roll-on and roll off characteristics of the cables such that the wing opens approximately as required. This pulley configuration is shown in Figure 18.
Figure 17: The two stage deployment.
Figure 18: The Root Hinge Roll-off and Roll-on pulley configuration
Figure 19 shows the ideal roll-on and roll off characteristic and the realized characteristic as a function of the root angle. The root hinge pulleys do not display any significant roll-on and roll-off in the first 90 degrees. Because there is no roll-on and roll-off, the next hinge lines are not able to open. The stack opens as a complete package. After about 90 degrees the length of the cables starts to change with the root hinge angle forcing the next hinge lines to follow. 30 30 30 | Document | AMS_2012.pdf | 122 |
109
Figure 19: The roll-on / roll-off characteristic.
The parameters in the ADAMS model are now tuned to the baseline Sentinel 1 design including the flexibility of the synchronization system. The model is used to predict the deployment characteristics under various thermal conditions and provides the following output: forces in the synchro-cables and control cables of the DSU, the torque provided by the DSU, shocks in the hinges when they lock, deployment time and path of the wing and the loads at the satellite interface. The ADAMS model is validated with the deployment test results of all relevant units and consequently the ADAMS results could be used to qualify the use of the DSU within the Sentinal-1solar array (verification of the occurring loads). It showed that the loads and shocks were within specification, the deployment time is within 88 and 284 seconds and deployment path of the Solar Array stays within the allowed deployment volume. Figure 20 shows the panel positions of the SA wing during deployment case 2, which provides maximum forces in the synchronization system.
Figure 20: The deployment path of Sentinel 1. Roll-off cable Roll-on cable | Document | AMS_2012.pdf | 123 |
110 Conclusion
The DSU for the Sentinel-1 solar array was successfully qualified and the flight models are produced (see Figure 21). Using engineering models proved to be essential to identify the deployment behavior of the Sentinel 1 wing and the required pulley diameter of the DSU. The qualification program showed that the alignment of the DSU needs to be controlled carefully to limit the friction of the control cables. The spring unit of the DSU design is over dimensioned for the Sentinel 1 SA; it can provide 1.5 times more torque than required, allowing application of the DSU in less favorable conditions / deployment systems. Representation of the Sentinel-1 deployment system (including DSU) in ADAMS allowed the actual wing deployment tests to be limited in both complexity and number of tests.
Figure 21: The DSU flight model. Acknowledgements
The presented work has been carried out at Dutch Space B.V. The authors would like to thank Mr. B. Busz, Mr. J. Cremers, Mr. P. Duyster and Mr. T. Konink of Dutch Space B.V. and Chris Verheul of Sayfield for providing data and support.
References
1. ESA Requirements and Standards Division.”Space engineering mechanisms.” ECSS-E-ST-33-01C, March 2009. 2. Ruag Space AG. “Datasheet Eddy Current Damper.” RUAG Space AG, Schaffhauserstrasse 580, CH-8052, Zurich, Switzerland. | Document | AMS_2012.pdf | 124 |
111 Ultra-low-weight Rotary Actuator for Operation on Mars and Pin Puller Mechanism Based on a No vel Shape Memory Alloy Technology
Nestor Nava*, Marcelo Collado*, Francisco Alvarez*, Ramiro Cabás*, Jose San Juan**, Sandro Patti*** and Jean-Michel Lautier***
Abstract A novel Shape Memory Alloy (SMA) has been developed as an alternative to currently available alloys. This material, called SMARQ, shows a higher working range of temperatures with respect to the SMA materials used until now. This temperature restriction is one of the most critical limitations of the current SMA devices for their use in space and other applications. A full characterization test campaign has been completed in order to obtain the main material properties and check its suitability for usage as an active material in space actuators. Results of this characterization test campaign have been presented in this work. This new alloy has been proposed for its use as actuators for space mechanisms. One application of SMA technology is an ultra-low-weight rotary actuator that has been developed for operation on Mars. The aim of this actuator is to provide an in-flight calibration system for the Dust Sensor instrument of the MEIGA-MetNet Mission that will perform airborne dust opacity measurements on the Mars surface. The total mass of the actuator is less than 9 grams (without control FPGA). A Qualification Model (QM) and a Flight Model (FM) will be presented in this work. The actuator presented is designed and qualified to withstand an impact inertia up to 2000 g and work at low temperatures (-90ºC) under vacuum conditions. Similarly, two versions of a Pin Puller mechanism working in the temperature range -30ºC to +125ºC have been designed and analyzed. Operative breadboard models of both devices were built and tested. The main characteristics of these devices as well as preliminary operating results will be shown in this work. The use of Shape Memory Alloys on the proposed actuators presents several advantages of lightweight, high force-to-weight ratio, and low volume. Introduction Shape Memory Alloys (SMA) can be defined as metals which, after an apparent plastic deformation in the martensitic phase, undergo a thermoelastic change in crystal structure when heated above its transformation temperature range, resulting in a recovery of the deformation that can be used to drive mechanisms [1]. SMA exhibit two properties, different than any other group of materials: the superelastic or pseudo-elastic effect and the shape memory effect [2]. When the material is at its high temperature phase, it can undergo large deformations by the action of an external stress and then instantly revert back to its original shape once the stress is removed. This behavior is known as pseudo-elasticity and it is due to the formation of stress-induced martensite. This martensite can withstand large deformations that can be completely recovered once the stress is removed [2]. When temperature is reduced, the material is transformed into twinned martensite, although if a mechanical stress is applied, the martensite structure is reoriented, producing a macroscopic deformation, apparently plastic. Nevertheless, when the material is heated, it changes to austenite, recovering its initial shape, as shown in Figure 1(a). The strain capabilities of this mechanism are usually limited to 7-8%. The martensitic transformation takes place in a temperature range that is one of the main parameters for the SMA alloys, and is called transition temperatures , as shown in Figure 1(b). The transformation occurs in the range defined by Ms (Martensite Start Temperature) and Mf (Martensite Finish Temperature). The reverse transformation (austenitic * Arquimea Ingeniería, S.L., Leganés, Spain ** Dpto. Física de la Materia Condensada, Universidad del País Vasco, Bilbao, Spain *** ESA ESTeC, Noordwijk, The Netherlands
Proceedings of the 41st Aerospace Mechanisms Symposium, Jet Propulsion Laboratory, May 16-18, 2012 | Document | AMS_2012.pdf | 125 |
112 transformation) takes place in the range between As (Austenite Start Temperature) and Af (Austenite Finish Temperature). Both effects are related to the thermoelastic martensitic transformation, which is a diffusionless reversible phase change characterized by a change in the crystal structure [3]. Thus, these characteristics allow SMA to be applicable for force and strain generation, in the case of shape memory, and for mechanical energy storage, in the case of pseudo-elasticity.
(a) (b) Figure 1. SMA characteristics: (a) microscopic diagram of shape memory effects; (b) martensite variation with temperature. The main advantage of the SMA technology for its use in actuators is its great strength in relation to its ultra-low weight, optimizing the mechanical work performed by the device with a minimum mass. Furthermore, this technology has the advantage of being immune to electromagnetic interference, its noiseless actuation, and that it does not require lubrication to work. The alloys currently available in the market, mainly NiTi based alloy, are limited in their operating temperatures. Arquimea suggests using a new SMA with further capabilities to overcome these temperature limitations. SMARQ material, a novel proprietary Shape Memory Alloy, is able to work in an extended temperature range, with transformation temperatures, tuneable during the manufacturing process, up to +180 ºC. We present in this work the results obtained from the SMA characterization. In order to perform the complete characterisation of the SMA material for its use as actuators in space environment, this document explains the way the tests have been carried out by considering the SMA material as the key element of a SMA actuator. This approach requires the space characterization of the complete actuator element in order to identify the performance of the different elements in the space environment. The main objective of these tests is to obtain technical information about the actuator performances. Special attention will be paid to the main advantage of the Arquimea’s SMA (SMARQ): its | Document | AMS_2012.pdf | 126 |
113 higher working range of temperatures with respect to the SMA materials used until now. This temperature restriction is one of the most critical limitations of the current SMA devices for their application in space missions and other markets. The present project is expected to demonstrate the capabilities of SMARQ to overcome this limitation. As above-mentioned, in order to perform the complete characterisation of the SMA material for its use in space environment, this document proposes to perform this characterization by considering the SMA material as the key element of a SMA actuator. A block diagram with all the constituent parts of a typical SMA device is shown in Figure 2. It must be taken into account that some of the blocks are basic blocks that must be part of any SMA device and others are optional, being or not a part of the device depending on the application requirements. The tests carried out as part as this work have been focused on the basic configuration of the SMA device, this is, the combination of the SMA Material and the mechanical and electrical interfaces. The characterization of the rest of the elements in space environment can be performed as independent blocks. The performance of the SMA material and the mechanical and electrical interfaces has been tested during the current work, in order to detect possible incompatibilities of the materials involved or mechanical limitations in the SMA actuators in space environment as well as to obtain information about the SMA actuator main parameters and working behavior. As an application of SMA technology, a Dust Sensor (DS) instrument will be presented in this work. The Dust Sensor instrument of the MEIGA MetNet Mission will perform airborne dust opacity measurements on the Mars surface. The Dust Sensor is designed as a lightweight device (41 g) that performs an active measurement, using back scattering to estimate the concentration of particles in the airborne dust. The unit integrates an infrared (IR) optical active detector (there is an IR emitter and an IR detector) for a spectrographic (discrimination in wavelength) measurement of dust in suspension.
Figure 2. Block diagram of Arquimea SMA Device. | Document | AMS_2012.pdf | 127 |
114 Figure 3. Qualification Model of the Dust Sensor
The emitter points to the airborne dust and the detector collects the scattering signal produced by the emitter light when interacting with particles similar in size to the light wavelength. Figure 3 shows the qualification model of the Dust Sensor device. The DS includes an in-flight calibration system based on a reflector stick that allows directing the emitter optical signal directly to the detector. The reflector stick is inserted in the optical path, when the actuator system is commanded to do so, by means of an actuator based on a Shape Memory Alloy. This in-flight calibration system is used to compensate the error in the entire DS acquisition chain. There is a weight budget of 40 grams for the Dust Sensor instrument, due to the small weight budget of the full MEIGA mission. This has greatly affected design decisions such as the lack of an enclosure, reduced support structures, limitation to the number and size of electronic components, and mainly in the actuator design. The mass budget for the actuator was less than 10 grams, which implied a challenge in design. The Dust Sensor device is controlled by the onboard computer of the MetNet Lander. Periodically, according to a preprogrammed schedule, the Dust Sensor device is powered on and a sequence of instructions is executed. Table 1 summarizes the main characteristics of the Dust Sensor device. A novel Pin Puller mechanism based on SMA technology will be presented. A Pin Puller is a mechanical device in which a pressure cartridge causes a pin or piston to retract inside the structure frame, usually against a side load. In the extended position, the pin or output shaft can be seen to be loaded by a compression spring. Table 1. Dust Sensor main characteristics.
Characteristic Value Notes Mass 41.2 grams Main dimensions 85x65x20 mm Power consumption 360 mW nominal 2250 mW peak consumption during 750 ms max. Voltage operation 5 V Communications 422 serial comms Command oriented instrument. Temperature accuracy ±1.5°C Reflector position accuracy ±7.5° Operational temperature range -90 °C to +25 °C Also in vacuum The pin remains firmly locked in this position due to mechanical components that block the stroke. Once actuated, however, the actuator drives specific mechanical components releasing the stroke and allowing the pin to retract under the force of the drive spring. The Pin Puller is reset by manually moving the pin back into the extended position. This is done by either pulling it out from the top or pushing it from the bottom. The Pin Puller acts as a trigger for deployment mechanisms. The inertia force of the Solar Array | Document | AMS_2012.pdf | 128 |
115 panels, antennas, booms, aperture door covers, etc. will be reacted through the release mechanism – through the Pin Puller in the shear direction across the pin and the housing to the structure. Pin pullers are normally used in pairs for redundancy. Applications include “Hold Down and Release” of numerous satellite deployable including solar panels, communication antennas, instrument cover doors, radiators, heat shields, tether experiments, and isolation systems. SMA Pin Pullers provide a number of advantages compared to traditional pyro devices: (1) There is no (or minimum) shock associated with SMA devices (note that there have been reported instances when pyro pullers have tripped relays when fired). (2) The number of safety personnel providing oversight during the installation of SMA devices would be greatly reduced. (3) Deployment tests can be performed repeatedly without having to remove/reinstall Pin Pullers. The product tree for the SMA Pin Puller mechanism, including its classification by trigger mechanism is shown in Figure 4.
Figure 4. Product tree for the SMA Pin Puller mechanism. A novel Shape Memory Alloy SMARQ is a fully European material technology and production processes, based on a low cost production procedure, which allows the manufacture of high quality products. Both material and production processes are currently being evaluated for use in space applications. In order to perform the complete characterization of the SMA material for its use in space environment, the SMA material (SMARQ) has been considered as the key element of a basic SMA actuator, consisting in the SMA material and the mechanical and electrical interfaces (Figure 2). The tests carried out as part as this work have been dedicated to the basic configuration of the SMA device. The tests have been focused on the actuator behavior when heated by means of an electrical current. The characterization of the rest of the elements – such as external heater elements, mechanisms, sensors or control electronics – in a space environment can be performed as independent bl ocks. A transitio n temperature (A s) of +145ºC has been selected for the test samples in order to satisfy the initial environment requirements of operation temperatures between -70ºC and +125ºC. | Document | AMS_2012.pdf | 129 |
116 SMARQ Characterization Test
A complete characterization test campaign for Arquimea’s SMA material (SMARQ) was planned and completed. This test campaign included several tests in order to demonstrate the material capabilities to be used as the trigger element in space mechanisms. The following blocks of tests were carried out: Strength characterization tests. o Maximum Strength Test o Reconditioning Force Test o Superelastic Test o Pull Force Performance Test Transition temperatures tests. Electrical activation tests. o Electrical activation at 22ºC. o Electrical activation at +125ºC. o Electrical activation at -70ºC. Lifetime tests at extreme thermal conditions. o Lifetime at +125ºC. o Lifetime at -70ºC. Material behavior tests in vacuum conditions. Assessment of material compatibility in space environment.
Strength characterization tests The scope of these tests is to obtain information about the force capabilities of the SMA actuator. A maximum strength test was carried out to determine the maximum applied stress the material can withstand without failure. A reconditioning force test has been performed in order to obtain information about the stress necessary to deform the SMA at its martensite phase and its behavior at different temperatures. An additional test, consisting of completing several load/unload cycles at different temperatures over the Austenite Finish Temperature (A f), in order to obtain the superelastic behavior of the material has been completed. Finally, a test was included to obtain information about the pull force performance (maximum force) of the material, by blocking the actuator movement during the phase transformation and measuring the generated force. A Tensile Test machine with oven has been used for the strength characterization. The equipment is able to perform universal tension tests at a controlled temperature, in the range between ambient and 200ºC. Transition temperatures test This test has been carried out to obtain data about the material transition temperatures (As, Af, Ms and Mf) for different samples. The non-operation temperature and the relationship between the strain and the temperature during transformation were obtained with this test. Arquimea’s Actuators Test Bench was used to complete this test, as shown in Figure 5(a). The equipment is specifically designed to test smart materials and actuators, allowing the execution of complete thermo-electro-mechanical characterization of both, in its actual working configuration. The bench can also control the actuation of the device using an electronic driver and it has an oven that allows heating the sample to a specific temperature.
Electrical activation tests An activation test has been performed at different temperatures in the working range (-70ºC, +125ºC). The material was loaded and externally activated by an electrical current. Measurements on strain, force, power consumption and generated force have been carried out for several samples, providing information to estimate the power consumption and the heating an d cooling times. Arquimea Actuators Test Bench of Figure 4(a) was used to complete this test at temperatures between ambient and +125ºC. For the cryogenic tests (-70ºC), a new test bench has been developed, as shown in Figure 4(b). A thermostatic bath was used to control the environment temperature of the sample at cryogenic temperatures down to -70ºC. The bench used a vertical measurement approach in order to allow the | Document | AMS_2012.pdf | 130 |
117 sample immersion in the thermal bath. Special attention was paid to the sample thermal isolation, in order to avoid the direct contact of the SMA with the thermostatic liquid, which would produce an important change in the convection conditions and therefore in the actuator behavior.
(a) (b) Figure 4. Devices used in the characterization campaign of SMAR: (a) Arquimea Actuators test bench; (b) test bench for cryogenic test. Lifetime tests The scope of this test is to obtain an initial approach to the lifetime capabilities of the material. The tests have been limited to probe that the material is able to complete 100 cycles at the worst case temperatures (-70ºC and +125ºC) without degradation of its performance. A lifetime of 100 cycles is a typical requirement for one-use space actuators, such as release and deployment actuators. The test bench of Figure 4(a) was used for the lifetime test at +125ºC, while the test bench of Figure 4(b) was used for -70ºC. Vacuum tests Due to the importance of the changes in power consumption and response times under vacuum conditions, a vacuum test was included in the SMA characterization test campaign. The objective of this test is to obtain technical information about the behavior of the SMARQ actuator in vacuum conditions at ambient temperature. A specific test bench has been developed for the vacuum tests. The equipment allows the strain and force measurement as well as the actuator electrical activation in vacuum conditions under constant loads. The test has been done controlling the pressure inside the vacuum chamber (~5·10 -3 mbar) and at ambient temperature (22ºC).
Test Results
Strength characterization tests The Maximum Strength Test was successfully completed. The material has shown its capabilities to withstand high loads, both in martensite and austenite phases. The material fractures at 500 MPa in its martensite state, showing an elongation over 12%. In the case of the test at higher temperatures, a stress of 400 MPa was applied without fracture of the sample. | Document | AMS_2012.pdf | 131 |
118 (a) (b) Figure 5. Results of strength characterization and transition temperature test: (a) reconditioning force – superelastic behavior; (b) SMARQ Transition Temperatures - Strain vs. Temperature. Actuation above 125ºC is shown.
These stress levels are far beyond the elongation and loads used in the actuators, so the material shows good properties for its application in terms of strength. The results show that a maximum superelastic strain around 10% would be feasible. During the Reconditioning Force Test, important information was collected about the force required to deform the material at different temperatures. The Superelastic Effect test has shown the typical behavior of superelasticity, Figure 5(a), with the existence of upper and lower load plateaus at different temperatures, where the deformation takes place with approximately constant stress. A maximum deformation of at least 8% could be achieved during superelastic cycles. Information about the Clausius-Clayperon curves was also obtained after this test. The Pull Force Performance Test has successfully been completed. The transformation has been observed and the stress levels obtained are high. The material output force capabilities are shown to be over 120MPa.
Transition temperatures test The test results of Figure 5(b) show that the material is able to work at environment temperatures over 125ºC, since the austenitic transformation starts, under the applied load conditions, around 145-155ºC. Besides, the material shows a martensitic transformation which finishes over 135ºC, so the material would be completely in martensite state along the whole environment temperature range (-70ºC, +125ºC). A thermal hysteresis of 15ºC has been obtained during the transformation and an 8% strain has been recovered. Electrical activation tests The test has shown the capabilities of the material to be electrically activated in the extreme temperatures in the working range, as well as in ambient temperature. This test shows that the material is able to work perfectly at these temperatures. The main differences between the tested temperatures are the energy necessary to complete the actuation, in terms of power and time, due to the higher difference between environment and transition temperatures. Information about the maximum non-firing current, activation currents, response times, and power consumption at each temperature have been obtained. The behavior of the wire during the electrical actuation was similar to the one produced with the oven in previous tests (transition temperatures test). A good stability and repeatability has been obtained along the cycles. Information about the wire resistance and its relationship with strain has been obtained. Moreover, an estimation of its relationship with the temperature has been carried out. | Document | AMS_2012.pdf | 132 |
119 Lifetime tests All the samples have successfully completed this test at both extreme temperatures. The values obtained during the 100 actuation cycles at each temperature for strain, stress, resistance and times show a high stability in the actuator behavior along its lifetime at the highest and lowest temperature in the working range. This shows that the present actuator technology is mature to be used in applications with these environment and lifetime conditions. The test has shown the material capabilities to satisfy the lifetime requirements in the current application. It should be taken into account that the samples have suffered a larger amount of cycles during the whole test campaign, so the actual lifetime obtained is >200 cycles. Figure 6(a) shows the high stability in the strain behavior along the cycles in the lifetime test at +125ºC. A comparative view for some of the cycles is shown in Figure 6(b). Future test will be conducted to verify the full capabilities of the technology in terms of lifetime. Tests will be completed in order to obtain the real lifetime limit for a 3.5% strain. Furthermore, other lifetime tests will be performed with different strain levels. (a) (b) Figure 6. Results of lifetime tests: (a) Lifetime (+125ºC) – Strain (%) vs. Time (s) (b) Detailed cycles: 1, 25, 50, 75 and 100. Vacuum test The cycles were successfully completed during this test. Complete actuations were achieved by using a power less than 1 W. The minimum required power to complete the actuation and the response time, are lower than the obtained in air, as could be expected. Finally, the material has not shown any problem related to the vacuum level during the test. More tests at higher vacuum levels will be done in the future in order to ensure the material works in a space environment. The main results obtained by SMARQ during the characterization test campaign in comparison to NiTi alloys are shown in Table 2. An excellent behavior in terms of operating temperatures was shown, with capabilities to operate over the limits of current SMA technologies. Good results in terms of reliability and lifetime were also obtained during this work. Successful results were also found during the vacuum tests. Further work is being carried out in order to obtain extra information about the material, especially related to lifetime and vacuum behavior. Research work for the material resistivity optimization are on course and based on the SMARQ heritage, a new generation of mechanisms based on this technology will be developed in the near future. | Document | AMS_2012.pdf | 133 |
120 Table 2. SMARQ Performances.
Property Units SMARQ NiTi Transformation Temperature (A f) °C > 150 (173°C) Max. 100 Non-actuating Temperature °C > 125 (As=152°C) Max. 80 Difference (A s - A f) °C 21 20 – 30 Thermal hysteresis °C 15 30 Max. Strain recovery (One way memory) % ~ 8 6 – 8 Maximum Strain (Superelasticity) % > 8 8 Recovery Stress (Martensite to Austenite) MPa ~120 300 – 600 (Max) Maximum Strength (Martensite) MPa ~500 800 – 1000 Elongation at Failure % 12 10 – 15 Stress Rate MPa/°C 2 4 – 20 Electrical Resistivity (Martensite) 10-6 Ω ·m 0.1 0.5 – 1 Power Consumption 10-6 W·m2/m 24.2 (min @ -70ºC, air convection)< 60 Response time s 4.5 s (Tested geom. @-70ºC, 4.5W, Af=173ºC) (*) 4.3 s (Test geom. @- 70ºC, 4.5W, Af=90ºC) Lifetime Cycles Tested > 100 (3.5%) (*) 100 (6%) 105 (2%) 107 (0.5%)
(*) No complete tests were performed to establish these parameters. Partial tests were carried out as a first approach. Neverth eless, the values are expected to be competitive with respect to NiTi.
Table 3. Rotary actuator main characteristics.
Characteristic Value Notes Mass < 9 grams without control FPGA Power consumption 2250 mW during 0.75 ms Voltage operation 5 V Torque 15.75 N/mm Rotary movement 40º Life >700 cycles Operational temperature range -90 °C to +25 °C Survival temperature range -90ºC to +70ºC Operation time in vacuum <750 ms 750 ms correspond to the worst case. This value depends on the initial temperature of the actuator. | Document | AMS_2012.pdf | 134 |
121 A Rotary Actuator for Dust Sensor
One application of SMA based actuators is the presented, which comprises a rotary mechanism actuated by a SMA Fiber, a position sensor, an electronic driver and a control algorithm. Table 3 presents main technical details of the rotary actuator integrated in the Dust Sensor device. The actuator integrated in the Dust Sensor is an ultra-light weight rotary actuator based on a Shape Memory Alloy (SMA) fiber. The SMA fiber contracts when heated beyond the characteristic transition temperature of the material. This contraction is used for generating a rotary movement of a stick reflector. The stand-by or relaxed position is reached thanks to a return spring. The actuator has an integrated rotary position sensor for control and characterization of the actuator movement. This rotary sensor is a double capacitance detector which gives a rough position value of the reflector. The consumption of this sensor is negligible and its weight is very low since part of the sensor is integrated in the PCB of the Dust Sensor. A limit switch (end of stroke) is included as a redundant element to switch off the activation of the DS. This switch will detect that the reflector stick has reached its final position, the signal will be interpreted by the FPGA and immediately the actuator will be powered off. The actuator hardware includes an electronic driver that provides the SMA fiber a power line of +5V, ~500 mA. For reliability reasons, the driver is commanded using a signal above 1 KHz from the control FPGA, neither a 5 V DC nor 0 V DC signal will activate the SMA. A control algorithm completes the actuator system. (a) (b) Figure 7. Initial (a) and further (b) position of the reflector stick. The algorithm running on the Dust Sensor FPGA commands the actuation receives information of temperature, rotary position and limits switch sensors and sends a feedback to the electronic driver. Figure 7 shows the initial and further position of the stick reflector during an actuation Qualification Campaign of Dust Sensor Qualification objective is to demonstrate that the Dust Sensor conforms to the requirements of the mission including margins. The following tests have been applied to the Qualification Model of the Dust Sensor: Thermal cycling and vacuum test from -90ºC to 70 ºC, 6 cycles. Vibration. QM DS was subjected to qualification levels of sinusoidal and random vibration. Bioburden reduction. Due to planetary protection requirements, a bioburden reduction process shall be applied to the DS unit. Humidity low-temperature verification test. A low-temperature verification test in a humidity atmosphere was carried out. | Document | AMS_2012.pdf | 135 |
122 Shock. QM of the DS was subjected to three different shock tests in each of the three axes. Qualification levels of the test are as following: o Axis X and Z: 500g in the form of sinusoidal half-wave with a duration of 2 ms. o Axis Y: 2000g in the form of sinusoidal half-wave with a duration of 15-20 ms. Test results The following qualification tests results are presented: Vibration. After vibration tests, a visual inspection, physic properties verification and several functional tests were carried out with the result of no anomalies or deviations detected. Consequently, the rotary actuator can survive the vibrations expected from travel, landing and take-off without any expected problem. Shock. Figure 8(a) shows the accelerations obtained during the shock test in X axis using the free fall machine. Similar results were obtained in Z axis. Figure 8(b) shows the accelerations obtained during the shock test in the Y axis using the pneumatic cannon. After the shock test, no anomalies were detected in the Dust Sensor, nor in the actuator subsystem. The levels in the Y axis during the shock test are slightly below 2000g. In the initial calibration tests using a mechanical dummy model, the values obtained arrived at 2000g. Unfortunately in the QM test this value was not reached due to the variability of the method applied. As seen in Figure 8(b), the levels reached have several peaks during approximately 17 ms of duration, which is more restrictive than the single sinusoidal half-wave requested. Even though the values obtained demonstrate that the DS can support shock values near 2000g: no damage or defects were observed in the mechanical dummy model (that reached 2000g), neither in the QM of the DS. Thermal cycling and vacuum test. Thermal vacuum cycling is a critical test for the rotary actuator subsystem due to the complex thermal processes that take place during the huge thermal variation in vacuum (from -90ºC to +70ºC). During the test, several data from the sensors of the DS were logged. Also, more than 100 actuations over all the temperature range were logged. Figure 9 presents an actuation capture during the TVAC test at -90ºC. As it can be seen, due to the absence of thermal convection losses, the SMA fiber heats very fast (less than 300 ms) and takes more time during the cold down (~1.25 s).
(a) (b) Figure 8. Results of shock test: (a) acceleration Vs Time obtained during the 500g shock test of the DS; (b) acceleration Vs Time obtained during the 2000g shock test of the DS. | Document | AMS_2012.pdf | 136 |
123 Figure 9. Position sensor output Vs time during an actuation at -90ºC.
Bioburden reduction. Bioburden reduction test verified that the actuator was not damaged after 50 hours at 111ºC. No anomalies or deviations were detected. The SMA fiber was not damaged and the rotary actuator worked perfectly. Humidity low-temperature verification test. During the verification test, inside the thermal chamber, the rotary actuator was completely operational. At low temperatures, convection losses were high enough to increase the actuation time to values as high as 3 seconds. Convection losses made the fiber not to contract completely and, as a consequence, perform a partial actuation (limited angle movement) at temperatures below -30 ºC. It must be noted that the rotary actuator is intended to be used in nearly vacuum atmosphere, with negligible thermal convection losses. As an additional test, in open atmosphere (49% of relative humidity) the critical parts of the rotary actuator were superficially completely frozen at -50ºC. Ice appeared all over the mechanism. Several actuations were performed in these conditions. The Dust Sensor rotary actuator was able to actuate in an atmosphere with vapor water within the temperature range without being blocked. The main lesson learned is that the complex thermal processes that take place along the whole temperature range of the mission are critical for a reliable and proper performance of this type of actuator. Here, the position sensor and end of stroke sensor included in the design play an important role for proper performance, and have allowed the completion of more than 700 cycles without failure during the qualification process. Future work includes the FM development and implementation in the MetNet platform. A Pin Puller Mechanism
A Pin Puller is another application for SMA-based actuators. Two different versions of the pin puller were designed during this activity. Figure 10(a) shows a CAD model of the first design of the Pin Puller V.1 . This design has been conceived to have a cylindrical shape in which length and diameter present similar dimensions, such as 80.0 mm of external diameter and 71.5 mm of length in un-actuated position (pin deployed). The mechanical parts have been designed to be made of an aluminum alloy. The commercial components have been modelled assuming their mechanical characteristics from the data sheets. The estimated weight of the whole structure, including commercial components and mechanical parts is 248.0 g. The Pin Puller mechanism is activated by a SMA actuator. Spheres support the pin at the initial position and a compression spring is loaded to perform the driving force once the release takes place. When the SMA actuates, a crown rotates allowing the spheres displacement, and thus the pin release and its movement to the actuated position (pin retracted). The mechanical design has been conceived in order to optimize the device’s weight, reduce the parts complexity, and achieve a suitable stiffness. The assembly ensures the alignment of components and compactness of the design. Note a M4 drill on the pin tip that allows installing a threaded tool that can be used to pull for the device reset. Flat surfaces both in the pin shaft and in the frame do not allow the undesired rotation of pin, preventing the mechanism to be released by an external force. | Document | AMS_2012.pdf | 137 |
124 Figure 10(b) shows an overview of the second design of Arquimea Pin Puller, named Pin Puller V.2. This design has been conceived in order to obtain a flat model structure. The device external diameter is 100.0 mm, the length is 50.9 mm with the pin retracted, and its estimated weight is 250 g. The Pin Puller V.2 presents a mechanism similar to the Pin Puller V.1 . Nevertheless, the V.2 pin is supported by rigid bars that keep it deployed during the un-actuated position. When a crown rotates by a SMA operation, the bars move, allowing the pin stroke by means of a compression spring. Pin Puller V.2 has also been conceived in order to have a low weight structure, no complex features, and a suitable stiffness. The reset is also planned to be done by pulling a M4 threaded tool located in the pin tip. The components alignment is also ensured in this Pin Puller, as result of the accurate design process, by using an internal ring that also ensures the design compactness, locating the supporting spheres. Note that Pin Puller V.2 has been designed to avoid any rotation of the pin. Mechanical and Thermal Analyses The 3D-CAD models of both versions of the Arquimea Pin Puller have been used to develop finite element analyses (FEA) of the structures. The goal of this analysis was to check the feasibility of the design as well as the components and assembly resistance. Two cases have been studied during the FEA, axial load condition and shear load condition. In particular, the axial load condition has been assumed as 180 N applied on the pin tip plus 1264 N applied by the stroke spring. The shear load condition has been assumed as 1000 N applied along a transversal axis of the pin. A factor of safety (FOS) of 1.25 has been suitably assumed both for incrementing the applied forces and for the numerical results in order to obtain a conservative design.
(a) (b) Figure 10. CAD models of Arquimea Pin Pullers: (a) V.1; (b) V.2. Figure 11(a) shows the results of the Von Misses stress from the FEA of Pin Puller V.1 structure for the second case of study (shear load). 450 N/mm2 has been computed as the maximum value (lower than the yield limit of the selected aluminum alloy). Figure 11(b) shows the stress results of the Pin Puller V.2 FEA. The maximum stress result has been computed as 290 N/mm2 (lower than the yield limit of the selected aluminum alloy too) during the second case of study (shear load). Successful results have also been obtained during the vibration analysis of both Pin Puller structures. The same set-ups used for the above-mentioned FEAs have been used for computing the limits of resonance, obtaining 2128.6 Hz for version 1 and 2350.4 Hz for version 2. These limits are above the expected operating frequency for both devices. Since mechanisms present difficulties during vacuum and extreme temperature conditions, tribology and thermal assessments have been done for both Pin Puller operations. Critical contact zones have been recognized in order to avoid high friction and adhesion. Certain solutions are proposed to resolve the problems, such as application of solid lubricant coating (MoS 2) on moving parts and designing the mechanical parts assuming specific features that reduce contact among components up to contact lines | Document | AMS_2012.pdf | 138 |
125 even contact points. Similarly, suitable tolerances have been assigned to moving part features in order to avoid any jamming produced by the material expansion or contraction at extreme temperatures. (a) (b) Figure 11. FEA results: (a) Pin Puller V.1 - shear load; (b) Pin Puller V.2 - shear load (right). Finally, the response time has been computed, as part of the actuator analysis, for both Pin Puller versions. In order to fulfil the environment requirements (working temperature between -30ºC and 125ºC), wires with Austenite Start Temperature (A s) of 150ºC have been considered. Therefore, FEA have been developed in order to obtain the temperature profile along the wire axis during operation and the actuator response time. The actuator maximum response time at worst case environment (-30ºC) has been estimated to be 4.3 s for version 1 and 2.8 s for version 2.
(a) (b) Figure 12. Built demonstrator of the both Arquimea Pin Puller versions: (a) devices of 500 N of stroke force; (b) devices of 100 N of stroke force.
Pin Puller as built Configuration Demonstration models (DM) of both versions, Figure 12, were constructed in order to test the mechanical device concepts and to debug the design. The DMs have been constructed in Alumide material by rapid manufacturing process. Alumide is a composite of plastic and aluminum that provides a suitable resistance to prototypes for preliminary operation test. Moreover, this material does not conduct electricity that ensures the electrical insulation of the structure. Since the device reset is planned to be made manually by pulling the pin, a reset tool has been designed and built too. The reset tool can be recognized in Figure 12 next to the Pin Puller DM. Finally, the DMs contain commercial components, such as screws, cables, crimps and SMA actuators, for completing a successful demonstrator device of a Hold-Down and Release Mechanism. Figure 12(a) shows the demonstrator models of a Pin Puller with a stroke force of 500N. Good functional performances were obtained from both models. Promising performances | Document | AMS_2012.pdf | 139 |
126 and some design concept improvements have been obtained for both versions of the device in this phase of the development. The designs presented in this work can be easily scalable to different sizes and requirements. Figure 12(b) shows the demonstrator model of a Pin Puller with 100 N of stroke force, in which an engineering model (EM) is currently under development.
Conclusions and Future Work
Arquimea has characterized the proposed SMA material (SMARQ), demonstrating the material capabilities to be used as a triggering actuator for space mechanisms. The material has shown a good behavior as Shape Memory Alloy. The project requirements in terms of operating temperatures, strain and force capabilities and lifetime have been satisfied. A rotary actuator included in the QM of the DS for environmental conditions requirements of the MEIGA-MetNet Mission has been presented. DS will perform airborne dust opacity measurements on the Mars surface, as can be confirmed from the qualification campaign results obtained. The rotary actuator operational temperature range from -90ºC to +25ºC was an initial challenge that has been completely reached thanks to a correct thermal design and a proper control algorithm. The temperature range obtained exceeds the operational range presented for other SMA-based actuator for a Mars surface application as stated in [4]. Two complete Pin Puller mechanisms were conceptually designed and analyzed, with promising results. Demonstration models for both devices were constructed and tested, obtaining a successful proof of concept of the mechanisms. The results of the work presented show that SMARQ technology is ready to be used in the development of EM, QM and FM actuators for future space missions. Other potential applications are the design of actuators for multi-cycle operation, for working in hard environments, such as cryogenic applications, or for applications where an ultra-light weight is required. Finally, the technology developed during this project can be applied to other industrial applications outside the space market.
References
1. Duerig, T.W., Melton, K.N., Stockel D., and Wayman C.M. (Ed.). Engineering Aspects of Shape Memory Alloys . London: Butterworth-Heineman, 1990. 2. Otsuka, K. and Wayman C.M. (Ed.). Shape Memory Materials . UK: Cambridge University Press, 1998. 3. Delaey, L. “Diffusionless Transformations”. Proceedings of Materials Science and Technology , (1991), Vol. 5, pp 339-404. 4. Fernández, D., Cabás, R. and Moreno, L. “Dust Wiper Mechanism for Operation in Mars”. Proceedings of 12th European Space Mechanisms & Tribology Symposium (ESMATS), (2007). 5. Jenkins P.P., Landis A.G. and Oberle L.G. “Materials Adherence experiment: Technology IECEC- 97339”. Proceedings of 32nd Intersociety Energy Conversion Engineering Conference, (1997). 6. Collado, M., Martínez, D., Álvarez, F., Cabás, R., Moreno, L. and López, F. “Evolution of SMA Rotary Actuators for Space Applications in Mars Environment”. Proceedings of Actuator 10 – International Conference and Exhibition on New Actuators and Drive Systems , (2010). | Document | AMS_2012.pdf | 140 |
127 Design and Performance of the Telescopic Tubular Mast
Mehran Mobrem and Chris Spier*
Abstract
The Telescopic Tubular Mast (TTM) has been under development at Astro Aerospace – Northrop Grumman Aerospace Systems for a number of years and has found several applications including deployment of the Large Sunshield on the James Webb Space Telescope. The TTM is composed of a number of large diameter tubes that are deployed and retracted using a Storable Tubular Extendable Mechanism (STEM). In order to study and evaluate the feasibility of utilizing the TTM concept for long boom applications a special design case (34.4-m long) was selected in 2005, which later was built and tested. This paper describes the design and analysis of this design case as well as testing performed. Special attention is given to the deployed stiffness and frequencies, which is a key requirement for space applications. Also, stabilization features required for the deployment are discussed and finally a feedback control system to drive the STEM deployer during the deployment is selected and the difficulties of controlling the system are discussed. Introduction The TTM is composed of a number of large-diameter thin-wall composite or metallic tubes as required to achieve a given deployed length, stowed envelope and structural characteristic. When deployed, the tube sections are latched together by multiple tapered pins to achieve a stable extended structure. These pins are withdrawn automatically to enable retraction. The Telescopic tubes are deployed and retracted by a version of Astro Aerospace’s STEM deployer. The STEM deployer has a long history of being used in space from the earliest small satellites in the 1960s as antennas and gravity gradient booms to the current GPS series of spacecraft. For the TTM application, the STEM has been evolved into a higher force actuator capable of more than 445 N (100 lb) deployment force. The TTM has to be deployed within a given time and it is desirable to have a constant velocity during the deployment or retraction. However, due to friction, latch up loads, and external tip loads, the velocity fluctuates. Also the friction and latch up forces fluctuate during the deployment/retraction depending on the individual segments. The STEM deployer inherently has a large dead band due to the winding/unwinding of its STEM element around the spool, therefore, each time the motor changes its direction it can not affect the deployment/retraction for a short period. Hence, the feedback control system should minimize the motor reversal during a given operation, e.g. deployment. Finally, it is desirable to limit the average velocity. TTM Top Level Information and Background The TTM test hardware is comprised of 17 telescoping tubes, excluding the fixed base tube. The tubes are optimized for stiffness, strength, and mass and are made of high modulus graphite composite with wall thicknesses ranging from 1.02 mm (0.040 in) to 0.38 mm (0.015 in). The tube closest to the fixed base is 31.8 cm (12.5 in) diameter, and tubes decrease in diameter by 1.3 cm (0.5 in) to the last tube at the tip which is 11.4 cm (4.5 in) diameter. The tubes are nested inside the fixed base tube and the nested group of tubes is deployed using a STEM deployer that is attached to the fixed base tube. The STEM deployer pushes the group of nested tubes Astro Aerospace – Northrop Grumman Aerospace Systems, Carpinteria, CA
Proceedings of the 41st Aerospace Mechanisms Symposium, Jet Propulsion Laboratory, May 16-18, 2012 | Document | AMS_2012.pdf | 141 |
128 from inside the fixed base tube. Latching features built into each tube allow the tubes to latch together as the system is deploying and creating a stable structure once fully deployed. The latching features are explained below. Stowed Design Features The base tube is designed to support the tubes and STEM deployer for a typical spacecraft application. There are two sets of attachment points one at the base and one in the side close to the tip. Also there are graphite load rings at the tube tips to provide a stiff structure for supporting the payload. The stowed configuration of the TTM is shown in Figure 1.
Figure 1. Stowed TTM
Deployment Design Features There are three major components in the design to stabilize the TTM during its deployment. Each tube has three rails along the tube axis that are spaced 120 degrees apart to stabilize the rotation about the boom axis during the deployment. These rails also have secondary effects to increase the deployed stiffness. The second major component is a secondary stiffening ring at the base about one diameter apart from the latch ring to stabilize the rotations about the two lateral axes. Finally, there is one centralizer in each tube to support the STEM element and stabilize/improve its buckling capability. These main features of each tube that help stabilize the TTM during deployment are highlighted in Figure 2. Figure 2. Tube Features Deployed Design Features The deployed TTM has nonlinear stiffness mainly due to local flexibility of the latching feature. The stiffness increases as the applied loads to the latch-pin increases. Several design features have been implemented to increase the deployed stiffness. A load ring has been added to each tube tip where the latch-pins are engaged to stabilizes the local deformation of the thin walled tubes. As discussed earlier these load rings are part of launch restraint to support the payload during the launch. Also, the three | Document | AMS_2012.pdf | 142 |
129 deployment rails along each tube add to the overall stiffness. Note that by pre-loading the latch-pins the stiffness would be increased and the non-linearity could be reduced. For many applications, there may be a type of tension load which could pre-load the TTM; however, if needed the STEM deployer could be utilized to do this function. STEM Deployer The drive system for the TTM is a STEM deployer powered by a brushless dc motor which pushes against the tip plate. The STEM consists of a “C” section of thin formed metal that is flattened so it can be rolled onto a spool for launch, as shown in Figure 3. Deployable booms in the STEM family are simple and extremely lightweight; they have been successfully deployed over 300 times in space without any known failures.
Figure 3. STEM device used for TTM deployment/retraction
The payload and package of stowed tube segments are pushed from the inside of the fixed external base tube by the STEM. When the package reaches the end of the fixed segment, the outer tube in the package latches to it. This tip deployment process repeats sequentially until all tubes are latched into place. The same sequence is reversed to retract. The innermost of the undeployed tubes is fixed to the tip of the STEM in order to stabilize the moving package of tubes.
To minimize the number of tubes, they are all the same length and are stowed coincident with each other. The latches fit in the annular gap between adjacent tubes in a stiffening ring at the lower end of each tube. The adjacent larger tube in turn necks down to a thin stiffening ring at the upper end. The stiffening ring helps to center and align the adjacent smaller tube and to lessen local deformations between the latched segments in bending. Tube Latching Small tapered pins are distributed circumferentially in the stiffening ring at the lower end of each tube. The pins are loaded radially outward by short springs to engage with tapered holes at the upper end of each larger adjacent tube, as shown in Figure 4. When stowed, the springs and pins are compressed by the interior surface of the adjacent larger tube. During Figure 4. Tapered pins used for latching | Document | AMS_2012.pdf | 143 |
130 deployment, the tips of the pins slide on the surface until they pop into the tapered seats to latch. Sequencing Since the tubes are stowed coincident to each other, each ring of compressed latch pins can engage the adjacent smaller ring with simple detents as shown in Figure 4. All the nested tubes are thus locked together so that they can be pushed as a package during deployment. When the latch ring in the outermost tube of the pack age locks it into deployed structure, t he detents retaining that tube to the moving package of tubes are released. The now smaller package of moving tubes continues without interruption. The male component of the detent on the interior end of the latch pin is conically shaped to make the latching function fail-safe. If one or more springs fail, the affected pin is forced out of the way by the female side of the detent, which acts as a ramp, as shown in Figure 4. Without the spring to preload the pin in the tapered receptacle, that pin cannot contribute to the deployed stiffness of the boom, however, deployment will not be impeded. Retraction To retract a given tube, its latch pins are pulled from engagement with the next larger tube by ramps in the next smaller tube. The ramps are hollowed out of the latch rings to engage conical rims at the male detent end of the latch pins, as shown in Figure 5. Latch pins are alternated with retraction ramps and detents in increments around the circumference of each ring. Each successive tube in the assembly is indexed by one such increment relative to its neighbors so that everything meshes properly. The length of the trough is controlled so that the detents will engage before the deployed tube is unlatched, as shown in the second inset of Figure 5.
Figure 5. Boom retraction sequence | Document | AMS_2012.pdf | 144 |
131 TTM Special Design Case
In order to study and evaluate the feasibility of utilizing the TTM concept for long boom applications, a special design case was selected in 2005 which was later built and tested. The key parameters and top level requirements for this design are: Deployed length of 34.3 m (1350 in) from bottom of base to tip Stowed length of 2.16 m (85 in) Mass is approximately 58 kg (128 lb) Deploy time within 14 minutes Stowed frequency should be greater than 35 Hz Deployed frequency should be greater than 0.1 Hz Mast should be able to retract by reversing motor Tip displacement during the deployment/retraction should stay within ±17.8 cm (7 in) Tube Properties There are 17 telescoping tubes, excluding the fixed base tube. Table 1 lists properties for the minimum and maximum diameter telescopic tubes, along with Tube 9 which is the tube in the middle between the base and the tip. Tube 1 is the tube closest to the base tube and is 31.8 cm (12.5 in) diameter, while Tube 17 is the last tube at the tip and is 11.4 cm (4.5 in) diameter. Listed in the table are the different stiffness properties for each of the tubes. EI is the bending stiffness of the tube, GJ is the torsional stiffness of the tube, and EA is the axial stiffness of the tube. Table 1. Tube Properties
Tube Tube Diameter (cm) Thickness (mm) EI (x103 N·m2) GJ (x103 N·m2) EA (x106 N) Mass/Length (kg/m) 1 (Max) 31.8 0.51 868.9 346.9 69.0 1.56 9 21.6 0.38 140.3 105.8 24.1 1.47 17 (Min) 11.4 1.02 81.1 41.9 49.6 1.93 Stowed Frequency Analysis A stowed finite element model of the TTM was created to compute the natural frequencies. The fundamental computed natural frequency is 37.88 Hz which exceeds the 35 Hz requirement. The first ten stowed natural frequencies are listed in Table 2 and the first three mode shapes are shown in Figure 6. Note that these frequency results have not been validated by a test.
Figure 6. First three modes in the stowed configuration | Document | AMS_2012.pdf | 145 |
132 Table 2. Stowed natural frequencies
Mode Frequency (Hz) Mode Frequency (Hz) 1 37.88 6 62.92 2 48.69 7 69.23 3 51.83 8 76.58 4 52.68 9 82.42 5 57.46 10 85.28
Deployed Stiffness Analysis and Test
A deployed nonlinear FEM was created to determine the deployed stiffness, natural frequencies, and its load capability. The nonlinear FEM was used to determine the deployed stiffness under different load conditions and was validated by test. Measurements were performed on the deployed TTM to characterize its stiffness under different loading conditions. The test setup consisted of the deployed mast offloaded with a total of 17 helium balloons. Each balloon was attached to the tip end of their respective tubes as shown in Figure 7. The lift of the individual balloons was adjusted to the mass of the supported tube sections. The tip load was applied with a motorized slide mechanism that had a load cell and position transducer attached.
Figure 7. Deployed stiffness test configuration
Baseline Tip Load – No Preload A baseline tip load test was performed with no preload. For this test, the mast was loaded in shear only at the tip to the levels shown in the plot in Figure 8. The blue curve in the figure shows the load-deflection characteristics of the mast with zero preload. The red line represents the predicted performance of the mast. The TTM measured stiffness correlates well with predictions, however it is lower at low amplitude and is higher at high amplitude compared to the predictions. | Document | AMS_2012.pdf | 146 |
133 Figure 8. Plot of tip displacement versus load for the no preload load case
Deployed Preload Condition The deflection of the fully deployed mast due to the application of a tip preload was also characterized. This type of loading generally relates to typical flexible solar panels or sun-shields under tension. The test configuration is shown in Figure 9. A typical 55.6-N (12.5-lb) load was applied in the direction shown at the tip and the resulting deflection was measured at the tip.
Figure 9. Deployed preload condition
The results of the preloaded test are shown in Figure 10. For this test, the mast was loaded with a 55.6-N (12.5-lb) tethered load, which creates an axial, shear, and moment load, and then an additional shear load was added at the tip to the levels shown in the plot. In this plot, zero deflection corresponds to the deflected shape under the tethered load, which was 13.8 cm (5.45 in). -4-3-2-101234 - 2 8 - 2 6 - 2 4 - 2 2 - 2 0 - 1 8 - 1 6 - 1 4 - 1 2 - 1 0 - 8 - 6 - 4 - 2 02468 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 2 6 2 8Load (N) Deflection (cm) | Document | AMS_2012.pdf | 147 |
134 Figure 11. First eight deployed modes The red line represents the predicted performance of the mast. This line shows the predicted tip deflection of 12.7 cm (5.0 in) under the tethered load versus the 13.8 cm (5.45 in) observed in the test. It was observed from this test that lower than predicted stiffness occurred at lower loads while higher than predicted stiffness occurred at higher loads. Figure 10. Plot of tip displacement versus load for th e preload condition Deployed Natural Frequency Analysis
A linear FEM was created from the nonlinear model with applied tethered loads to compute the natural frequencies. The fundamental computed natural frequency is 0.1 Hz which meets the 0.1 Hz requirement. Note the mode shape is orthogonal to the plane of load application where the boom stiffness is at minimum stiffness region. The first eight deployed natural frequencies are listed in Table 3 and the mode shapes are shown in Figure 11. Note that these frequency results have not been validated by a modal test. A detailed view of the deployed FEM is shown in Figure 12.
-3-2-101234 -35 -30 -25 -20 -15 -10 -5 0 5 10 15 20 25Load (N) Deflection (cm) Figure 12. Deployed FEM | Document | AMS_2012.pdf | 148 |
135 Table 3. Deployed natural frequencies
Mode Frequency (Hz) Mode Frequency (Hz) 1 0.10 5 1.56 2 0.15 6 2.09 3 0.56 7 3.09 4 0.77 8 4.09
Deployment Characteristics and Control The goal was that the TTM should be deployed within 15 minutes. It is desirable to have a constant velocity during the deployment or retraction; however, due to friction, external tip loads, and latch up loads the velocity fluctuates. Also, the friction and latch up forces fluctuate during the deployment/retraction depending on the individual segments. A typical resistive force during a segment latch up is show in Figure 13.
Figure 13. Typical latch loads during deployment
The drive system is a STEM deployer powered by either a brushless dc motor or a stepper motor which pushes against the tip plate. The STEM deployer inherently has large dead band due to the winding/unwinding of its STEM element around the spool, therefore each time the motor changes its direction it can not affect the deployment/retraction for while. Hence the feedback control system (required for brushless dc motor application) should minimize the motor reversal during the deployment operation. Finally, it is desirable to limit the average velocity. -0.20.00.20.40.60.81.01.2 05 0 1 0 0 1 5 0Normalized Load Time (sec) | Document | AMS_2012.pdf | 149 |
136 Deployment Drive Feedback Control System To achieve the deployment criteria, a feedback control system containing the following is considered (see Figure 14): Position loop Velocity loop inside the position loop Feed forward time dependent velocity profile - trapezoidal Time dependent reference position – integral of velocity profile Both velocity and position loops should be: o Stable for minimum inertia due to the backlash (motor inertia only) as well as maximum inertia o Open loop frequency bands should be less than 70% of the structural natural frequency o Nonlinear gains used in both position and velocity loops o Gain and phase margins should be greater than 10 dB and 40 degrees respectively Current limit is required to prevent high deployment force
Figure 14. Control system block diagram In Figure 14, V tr is the translational drive reference velocity, Td is the approximate differentiator time constant, L is the motor inductance, R is the motor resistance, τe is equal to L/R, kT is the motor torque constant, and kb is the back emf constant. The gains for the control system are: k p is the position loop gain, kv is the velocity loop gain, and kR is the resolver gain. It is desirable to have nonlinear gains in both the position and velocity loops in order to prevent the motor reversal caused by back winding in the STEM. A possible nonlinear control gain that can be implemented is a piece-wise linear gain that is high gain for lag and low gain for lead. The TTM is a time variant system and, because of this, there are non-conservative forces during the deployment. These forces become significant as the deployment velocity and acceleration increases | Document | AMS_2012.pdf | 150 |
137 [2,4]. A similar feedback control system was designed and tested for the Space Station Freedom Mobile Transporter [2,3]. For the Mobile Transporter, an avionics control breadboard simulator was designed to simulate and verify the operation of the Mobile Transporter control system. The following dynamic models are used to derive the TTM system state space matrix [A,B,C,D]. Note that during deployment the length of the TTM is changing causing a change in natural frequencies. Primary Dynamic Model The primary dynamic model used for the design of the control system is shown in Figure 15. This model represents the deployment and axial vibration of the TTM. The moving mass M is the mass of the deploying tubes plus the mass of the payload and fresistive are loads due to friction, tube latching, and any external loads such as sun-shields tension and/ or space craft attitude control. For the STEM deployer, k STEM is the axial stiffness of the partially deployed STEM and varies by length, θm and θs are the rotation of the motor and spooler, respectively, Jm and J s are the inertias of the motor and spooler, respectively, and n is the gear ratio.
Figure 15. Dynamic model for control system The equations of motion for the STEM deployer and the system of deploying tubes shown in Figure 15 are resistive STEM f x x k x M 0 1 1 (1) The axial natural frequency changes as the STEM deploys and the boom length changes. A plot showing the change in axial natural frequency as the system deploys and the length increases and moving mass of the tubes decreases is shown in Figure 16. | Document | AMS_2012.pdf | 151 |
138
Figure 16. Plot of axial frequency versus deployed tube bays
Secondary Dynamic Model The secondary dynamic model used for the design of the control system represents the lateral vibrations of the TTM during the deployment. This can be represented by an axially moving cantilever beam with a tip mass as shown in Figure 17.
Figure 17. Cantilever beam with tip mass and time dependent length
The equation of motion for the system shown in Figure 17 for a cantilever beam whose length changes with time is given by [2] 0) (6857 . 0) (8 . 4 028 . 1 ) (12 35334 ) (35332 3 ut Lmat LM m t LEImu M t mLtip tip (2)
where u is the velocity of the moving beam, a is the acceleration of the moving beam, EI is the bending stiffness of the beam, m is the mass of the beam, L(t) is the instantaneous length of the beam, Mtip is the tip mass and η is the normal coordinate representing the lateral displacement y. Note that equation (2) has non-conservative forces which could become important depending on the deployment velocity and acceleration. 00.20.40.60.811.2 0123456789 1 0 1 1 1 2 1 3 1 4 1 5 1 6 1 7Normalized Frequency Deployed Bay | Document | AMS_2012.pdf | 152 |
139 If the velocity and accelerations are small, equation (2) reduces to
0 ) (124 ) (3533 3 t LEIM t mLtip (3)
Control System Simulation A simulation of the control system was performed using Simulink. A resistive force representative of the one shown in Figure 13 is used to simulate latching loads. Non-linear control gains are used for the position and velocity loop gains and a deadband of ±7.6 cm (3 in) is used. Simulation results for the TTM deployment are shown in Figures 18 and 19. Figure 18 is a plot of velocity versus time and includes deadband in the system. Figure 19 is also a plot of velocity versus time, but does not include deadband in the system. In both plots, time is in seconds and velocity is in meters per second.
Figure 18. Velocity versus time plot with deadband
Figure 19. Velocity versus time plot with no deadband | Document | AMS_2012.pdf | 153 |
140 Summary
A deployable and retractable telescoping boom that contains large-diameter, thin-walled composite tubes has been developed that is capable of high deployed stiffness. Tapered pins are used to control the deployment and retraction between the tubular sections and create a stable extended structure when deployed. The latch design and STEM work together to eliminate the need for segments to overlap when deployed giving a lightweight design. The key parameters and requirements for the TTM study are given, and it is shown by analysis and test that the requirements are met. Frequency analysis on the deployed and stowed configurations was performed using FEA software to show that the frequency requirements are met. Testing was also performed on the deployed TTM to characterize the stiffness under different loading conditions. In all the load conditions, lower than predicted stiffness occurred at lower bending loads and higher than predicted stiffness occurred at higher loads. Control of the deployment drive for the TTM presents several challenges due to the design of the STEM actuator and the latching loads during deployment. To analyze the system, a dynamic model was developed that includes the deploying tubes and the STEM motor and spooler. The equations of motion for the system were derived and a feedback control system was designed. The work done to date and the progress made on the TTM system are presented here. References 1. Thomson, M.W. “Deployable and Retractable Telescoping Tubular Structure Development.” Proceedings of the 28 th Aerospace Mechanisms Symposium (May 1994). 2. Mobrem, Mehran and Mark Thomson. “Control Design of Space Station Mobile Transporter with Multiple Constraints.” First Joint U.S./Japan Conference on Adaptive Structures , (November 1990), pp. 87-116. 3. Mobrem, M., Paden, B., Bayo, E., Devasia, S., “Optimal Output-Trajectory Tracking: Application to Mobile Transporter Avionic Breadboard.,” AIAA Guidance, Navigation, and Control Conference and Exhibit, (August 2000)
4. Taborrok, B., Leech, C.M., Kim Y.I., “On the Dynamics of an Axially Moving Beam,” Journal of the Franklin Institute , Vol. 297, No. 3, (1974), pp. 201-220. | Document | AMS_2012.pdf | 154 |
141 Development of Variable Reluctance Resolver for Position Feedback
Greg Leibovich* and Sara Senanian*
Abstract
The variable reluctance resolver (VRR) is commonly used in the automotive industry as a rotor position sensor. Its use in the Aerospace and Defense Industries has been limited due to inadequate accuracy performance. However, recent interest in the Aerospace Industry for VRRs as a viable alternative to conventional resolvers has driven the development of higher accuracy VRRs for use in angular position feedback applications. This paper presents the progression in the development of the VRR and provides information on the lessons learned during the completion of the first generation of single speed and multi-speed engineering units of VRRs. This first generation of engineering units was evaluated to identify potential design and manufacturing process improvements. Incorporating the lessons learned from the development of the first generation units resulted in an improvement in performance and manufacturability in the second generation of VRRs, making it a desirable option for future use in commutation of brushless DC motors. Figure 1. Single-speed (left) and multi-speed (right) variable reluctance resolvers Introduction Resolvers are rotary motion feedback sensors used to provide angular position detection in a wide variety of environments; space, defense, automotive and oil exploration. For decades, conventional resolvers have been used for high precision mechanism position feedback and commutation in a range of adjustable speed drives; vector control induction motor, switched reluctance motor and brushless DC motor drives. Besides resolvers, optical incremental encoders and potentiometers offer an alternative solution for position sensing by producing precise incremental position information. However, optical encoders are susceptible to vibration, debris and high temperatures. Potentiometers are easy to use; however the potentiometer contains an electrically conductive wiper that slides across a fixed resistive element causing the possibility of considerable wear. Hence the life cycle of a potentiometer is limited and its sensitivity to vibration is of significance. Resolvers do not depend on moving electrical contacts for signal reliability and do not exhibit significant aging or changes to performance due to extreme temperatures or vibration. * Ducommun LaBarge Technologies, Carson CA
Proceedings of the 41st Aerospace Mechanisms Symposium, Jet Propulsion Laboratory, May 16-18, 2012 | Document | AMS_2012.pdf | 155 |
142 In addition, conventional resolvers transmit high accuracy angular data electrically in the thousands of a degree (arc-seconds) and consist of a stator, a rotor and a transformer assembly; each with its own winding as seen in Figure 2. Nonetheless, the conventional resolvers are intricate to manufacture and therefore labor intense and costly. Furthermore, in larger conventional resolvers, the transformers have been observed to delaminate and are difficult to rework or replace. The high cost and manufacturing complexity of conventional resolvers lead to the development of variable reluctance resolvers . Unlike a conventional resolver, the VRR has no need for rotary transformer and it consists of a simpler design. All windings are located on its stator assembly whereas in conventional resolvers the stator core, rotor core, stator transformer and rotor transformer each contain windings as labeled in Figure 2. Figure 2. Conventional resolver with transformer Furthermore, the study and experiments conducted by the Ducommun Team resulted in an improvement in the accuracy of the VRR making them competitive to convention resolvers particularly for the commutation of brushless DC motor drives in the Space and Defense Industries. This paper documents the development and qualification of the VRR for use in commutation of brushless DC motors that minimizes envelope size, mass and increases reliability. The purpose of development of VRR 1. Provide resolver architecture for continuous operation without rotary transformer 2. Reduce amount of magnetic components and windings 3. Decrease the manufacturing complexity and fabrication cost 4. Reduced envelope size and mass The VRR was developed to meet the following accuracy requirements: 1. Single speed accuracy to be less than 30 arc-minutes 2. Multi-speed accuracy less than 6 arc-minutes Background
The VRR consists of stator core and a rotor core and is unique in that its rotor core is slot-less with no windings. Unlike conventional resolvers, the VRR is based on variable reluctance between the stator and rotor segments. The voltages of the output windings of the resolver vary as sine and cosine functions of the rotor angular position. The VRR is structured to vary the reluctance of the magnetic path between the stator and rotor core in accordance with equation (1) through the change in the length of the air gap ( l) or the area of the air gap ( A). From equations (2) and (3) it is easily determined that the variation in reluctance causes a change in the induced voltage in the sense windings. Therefore, varying the air gap | Document | AMS_2012.pdf | 156 |
143 or coupling area will cause a variation in the reluctance and consequently cause a change in the voltage outputs. ൌ ఓ (1)
Φൌ ௧ ௨௧ൌ ே ࣬2)
ௗథ ௗ௧ (3) Where ࣬is the magnetic reluctance; ݈ is the length of the air gap; ܣis the area of the air gap; Φ is the magnetic flux and ݁is the induced voltage. For a multi-pole configuration, as seen in Figure 3, the structure is such that the variation in the length of the air gap between the stator and the rotor produces sinusoidal variations in the output voltages. Figure 3. Structure of multi-speed variable reluctance resolver For absolute position feedback a single speed variable reluctance resolver is used, which consists of a couple-pole configuration. The configuration of a couple-pole VRR varies the reluctance through the variation in the coupling area between the stator and rotor core segments. A unique circular rotor core is supported by two non-magnetic members, as seen in Figure 4. The rotor core provides a diametrical diagonal magnetic path from upper half of stator core to the lower half of stator core as illustrated below. Figure 4. Structure of single-speed VRR Development of Single Speed Variable Reluctance Resolver Description of first generation single speed VRR The architecture of the first generation single-speed VRR, as seen in Figure 5, consisted of two laminated stator cores and a single laminated rotor core. Each stator core, made of high permeability laminations, were stacked, bonded and insulated. | Document | AMS_2012.pdf | 157 |
144 Figure 5. First generation single speed VRR
The two stator cores were then inserted into the back iron and aligned and then bonded to the back iron. The excitation winding was placed between the two laminated stator half cores diametrically, as shown in Figure 6, and the sense winding was inserted through both cores axially. Figure 6. Architecture of the single speed VRR
The rotor core consisted of high permeability laminations, which were stacked and placed on a temporary hub against spacers on either side of the lamination stack and were bonded. The rotor core was designed as a laminated core in order to minimize rotor reluctance since a solid rotor has high reluctance due to eddy current losses. The laminated rotor core produced a high enough transformation ratio of the output voltage to the input voltage. Shortcomings of first generation single speed VRR During the manufacturing process and the functional testing, the following were issues which called for design improvements: The manufacturing process for the rotor core was extensive and consisted of several machining steps. After the rotor laminations were bonded lamination-to-lamination and lamination-tospacers, the rotor assembly was machined on the outer diameter (O.D.), then removed from fixture and ground on the inner diameter (I.D.). Finally the rotor assembly was bonded to a hub and its O.D. ground to final dimension to improve concentricity. The several machining processes introduced various stresses to the assembly and jeopardized the bonding adhesive; which introduced de-laminations of the rotor core. The thickness of the insulation on the stator cores was inconsistent which contributed to the tolerance stack-up of the stator cores in the stator assembly. Furthermore, the insulation used which was Scotchcast Electrical Resin #260 (3M) was observed to chip off the stator teeth during testing and handling of the VRR.
During accuracy testing, the accuracy of the single speed VRR was initially found to be approximately 85 arc-minutes as highlighted in Table I, which was much greater than the design | Document | AMS_2012.pdf | 158 |
145 -20246810 0 100 200 300 400Arc minutes DegreesRadial Misalignment Accuracy Test fo r Single Speed VRR offset by 0.001" Delta Accuracy Readingaccuracy requirement of 30 arc-minutes. Following the initial accuracy test, various tests were conducted for different parameters to locate which were the greatest contributors to the accuracy of the resolver. The numerous tests included recordings of the signal outputs of the resolver from radial misalignment testing, axial misalignment testing, stator core alignment and repeatability testing. It was established that the stator core alignment had the greatest contribution to the accuracy. See Table I for accuracy readings obtained after the stator core alignment was improved with a significant increase in accuracy from 85 arc-minutes to approximately 48 arc-minutes. For rotor misalignment, the radial misalignment was found to be of higher contribution to the accuracy than axial misalignment with a deviation in the accuracy reading of approximately 9 arc-minutes at the off-set position of 25 µm (0.001 in) as seen in Figure 7. Table I –Accuracy Test Results of Single Speed VRR Angle (degree) 0 20 60 100 140 180 220 260 300 340 360 Initial Accuracy Test (arc-min) 0.0 -5.9 -17.8 17.5 44.2 84.4 -40.4 42.6 -48.0 -9.6 0.0 Accuracy after Stator Core Alignment (arc-min) -0.3 -6.6 -13.1 -2.4 20.1 43.6 47.6 36.3 24.3 7.7 0.3 2nd Gen VRR Accuracy Test (arc-min) 0.0 -9.0 -14.9 -3.4 1.6 -5.3 -9.1 -7.7 6.3 5.6 0.0
Figure 7. Radial misalignment test result Lessons learned from the first generation Single Speed VRR The observations made and the lessons learned from the manufacturing and initial testing of the single speed VRR were the following: Excessive machining of the rotor core assembly was causing de-lamination of the rotor core and needed to be dimensioned such as to minimize the machining required to achieve final dimensions. Based on the stator core alignment test, it became apparent that the variation in the insulation thickness affected the accuracy significantly; therefore it was found that masking the stator core surface at the outer edge prior to applying insulation was critical. Lastly, a final grinding of the O.D. of the rotor core assembly was observed to be necessary to increase concentricity since the accuracy was found to be greatly affected by radial misalignment. Design improvements for the second generation single speed VRR The lessons learned from the first generation were implemented in the next generation single speed VRR through the following design modifications: | Document | AMS_2012.pdf | 159 |
146 The stator and rotor cores were re-designed with the design objective of a solid rotor core in order to eliminate any possibility of de-lamination. This was obtained by increasing the slot area of the stator cores to increase the coil turns in order to compensate for the eddy current losses in the solid rotor and to achieve the required transformation ratio. The stator cores were masked at the edges and were electrostatic powder coated using the Scotchcast Electrical resin 5133 allowing for a thinner and more consistent insulation thickness with no chipping observed. All machining of the rotor core was minimized by machining the solid rotor core to final dimensions less the O.D. which once bonded to the hub would undergo a minimal O.D. grind to improve concentricity. The second generation single speed VRR showed significant enhancement in performance once the design improvements had been implemented. Testing of the second generation VRR produced an accuracy reading of approximately 15 arc-minutes as listed in Table I. Comparing to the initial reading of 85 arc-minutes it was a great achievement. Development of Multi Speed Variable Reluctance Resolver
Description of first generation multi speed VRR The multi-speed VRR was designed and built with a single laminated stator core with specifically designed tooth profile for optimized sensitivity of variable air gap and a slot-less laminated rotor core. The resolver was developed with an axial length of 6.35 mm (0.250 in) offering a compact design, which resulted in a 30% reduction in mass compared to that of a conventional resolver of similar O.D. The excitation windings and the sense windings were placed on the stator core alone. Its air gap was varied such to produce a sinusoidal change in the magnitude of the output voltages of the two sense windings. Observations and test results of the multi speed VRR The performance of the multi-speed VRR was exceptional and produced accuracy test results much higher than required of the design. The short coming of the initial design was limited to the insulation material Scotchcast resin #260 which was chipping as it had in the single-speed VRR. Accuracy testing was performed and the resolver accuracy was measured in every 1 degree increments from 0 – 360 degrees and its results are summarized in Table II. The highest reading was found at 350 degrees at a value of approximately 35 arc-seconds which is listed and highlighted in the Table. Comparing to the design requirement of 360 arc-seconds (6 arc-minutes) the multi-speed VRR performed much greater than anticipated. Table II –Accuracy Test Results of Multi-speed VRR Angle (degree) 0 20 60 100 140 180 220 260 300 340 350 360 Accuracy Test (arc-sec) 0.0 36.7 0.1 -21.5 27.9 6.5 -20.5 25.5 3.9 -24.9 35.2 0.0
Conclusion / Summary The Ducommun team has developed Variable Reluctance Resolvers with sufficient accuracies to be in used in the Aerospace and Defense Industries in angular position feedback applications. The single-speed and multi-speed designs have been fully qualified and were shown to meet internal as well as external customer requirements. The latest development of the variable reluctance resolvers has made them competitive to conventional resolvers and offers an alternative solution in the application of brushless DC motor commutation. | Document | AMS_2012.pdf | 160 |
147 FeF 3 Catalytic Influence on PFPE Lubricant’s Lifetime under Loaded Conditions
Lionel Gaillard*, Antoine Mariot**, Catalin Fotea* and Roland Holzbauer+
Abstract
A Perfluoropolyalkylether (PFPE) oil was heated in an inert atmosphere under dynamic temperature control. The temperature profile allowed precise determination of the degradation onset temperature. The degradation temperature appears to be lower than literature values under the conditions used for this investigation. Other samples of PFPE oil were later tested with a Pin on Disk Tribometer under different temperature and loading conditions.
Introduction
While PFPE lubricants are widely used in mechanisms designed for aerospace industry because of their numerous advantages, it is known that they degrade quickly and heavily under boundary lubrication conditions, especially when in contact with 440C or 52100 stainless steels. Numerous authors (1) (2) (3) (4) (5) (6) in the past decades suggested that degradation of PFPE oils under boundary lubrication conditions was catalyzed by the Lewis acid FeF 3 (Iron fluoride III). Some authors (3) even proposed the degradation mechanism was autocatalytic, which means it could continue for a while after the mechanical stress ceases. In 1991, David J. Carré (2) proved indeed that PFPE does degrade in the presence of FeF 3 by reacting a branched PFPE and FeF 3 in a nickel-lined autoclave under inert atmosphere at high temperatures. The reaction is triggered in this instance at a temperature approximately 30°C below thermal degradation; however, high mechanical stress could have contributed to the high temperature excursions (7). The recent experiment, on the other hand, has dealt specifically with the degradation temperatures and more importantly to precisely assess the triggering temperature for PFPE degradation. This is the first experiment that combined the precise temperature control for the heating ramp by using a state of the art Netzsch Simultaneous Thermal Analyser (STA) while monitoring the discrete changes that occur within the PFPE/FeF 3 mixture using Fourier Transform Infra-Red (FTIR) and a mass spectrometer. Furthermore, the experiment would provide further insight into the autocatalytic mechanisms as suggested by (3).
Experiment
The Netzsch STA consists of a dynamic temperature controlled furnace coupled with a Fourier Transform Infrared Spectroscopy and Mass Spectrometry (Figure 1 and Figure 2). The instrument is vacuum tight by design, the furnace is evacuated first at pressures of ca. 10 -2 mbar and is continuously purged with N 2 gas therefore the pressure inside the furnace is atmospheric during continuous operation. * European Space Agency, Noordwijk, The Netherlands ** École Nationale Supérieure d'Arts et Métiers ParisTech, Lille, France + Aerospace & Advanced Composites GmbH, Seibersdorf, Austria
Proceedings of the 41st Aerospace Mechanisms Symposium, Jet Propulsion Laboratory, May 16-18, 2012 | Document | AMS_2012.pdf | 161 |
148 Figure 1. Furnace (left) and mass spectrometer (right).
Figure 2. FTIR facility. The fact that we were working at a more or less ambient pressure should not be an issue since even in high vacuum both FeF 3 and PFPE oil are expected to be present as a condensed form. In this case, the pressure dependence of the Arrhenius equation is low. Approximately 5 mg of FeF 3 powder were mixed with ca. 100 mg of Fomblin Z25 PFPE lubricant in an aluminum crucible. Fomblin Z25 is a base oil used in various Maplub and Braycote greases. These quantities of FeF 3 were intentionally high to ensure that iron fluoride would be in excess. It was feared that the reaction rate with “realistic” amounts of FeF3 (which is supposed to form as a near-monomolecular layer on real-life mechanisms (3)) would be too slow to be measured during a reasonable timespan. A “witness” cup containing only Z25 oil was also prepared. The temperature profiles used for this exercise are shown in (Figure 3). | Document | AMS_2012.pdf | 162 |
149 Figure 3. Temperature profiles. At the time of writing this paper, the following tests were completed: 1. Neat Fomblin Z25 was heated to 1300°C using a 10 K/min ramp to assess the degradation temperature of the lubricant. 2. A mixture of Fomblin Z25 and FeF 3 was heated to 300° using a 10 K/min ramp. 3. A mixture of Fomblin Z25 and FeF 3 was heated to 300°C using a 10 K/min ramp and cooled down to 40°C using a -10 K/min ramp. 4. A mixture of Fomblin Z25 and FeF 3 was heated to 300°C using a 10 K/min ramp and cooled down to 40°C using a -10 K/min ramp, then maintained at 40°C for one hour.
The FeF 3 has never been clearly identified in an actual mechanism or test apparatus, even if its presence is very likely. Consequently, additional tests on a tribometer were run. The aim was to observe the gaseous compounds coming out of the degradation reaction. Although it would not be a definitive proof, finding the same degradation products and behavior as in the STA facility would greatly increase the 0 Time (hours) Temperature (°C) 10 K/min 40 1300 1 2 0 Time (min) Temperature (°C) 10 K/min 40 300 10 25 5 15 20 0 Time (min) Temperature (°C) 10 K/min 40 300 10 50 30 40 20 - 10 K/min 0 Time (hours) Temperature (°C) 10 K/min 40 300 2 1 - 10 K/min | Document | AMS_2012.pdf | 163 |
150 likeliness of the hypothesis. A Pin-on-Disc (PoD) tribometer was used in a high-vacuum chamber under temperature control, along with mass spectrometer analysis.
Results and discussion
The heating of neat lubricant to 1300°C determined the thermal degradation onset temperature of the Fomblin Z25 oil. Very precise onset points for the lubricant mass loss and the FTIR Gram-Schmidt reconstruction readings were observed. Figure 4 presents the temperature profile (red curve), the total mass loss of the oil (green curve), the differential scanning calorimetry (DSC) reading (blue curve), and the FTIR Gram-Schmidt reconstruction (black curve). The thermal degradation occurs at ca. 400°C.
Figure 4. PFPE run to 1300°C. | Document | AMS_2012.pdf | 164 |
151 Figure 5. Z25 sublimation test under high vacuum. Observation: the experiment was run at an ambient pressure, which means that under vacuum, sublimation will occur before the oil reaches 400°C. This was shown by a sublimation test run by AAC GmbH (Figure 5). Given those results, it was not necessary to go above the thermal degradation temperature, which led us to choose a 300°C maximum limit. The second (PFPE and FeF 3 mixture heated to 300°C using a 10 K/min ramp) and third test (mixture of Fomblin Z25 and FeF 3 heated to 300°C using a 10 K/min ramp and cooled down to 40°C using a -10 K/min ramp) showed that some degradation was occurring at a much lower temperature. Figure 6 shows an onset point for the mass loss (green curve), FTIR Gram-Schmidt reconstruction (black curve) and mass spectrometer readings (dotted curves). Thus, the degradation appears to be triggered at ca. 220°C. Figure 6 presents the different gaseous products detected by the mass spectrometer as follows: the main peaks were observed at atomic mass units (amu) 19, 44, 47, 66 and 69. The 19 amu trace can be attributed either to electron-stimulated desorption (along with 1 and 16 peaks, not shown) (8) or to fluorine radicals coming out. The other traces are 44 amu: CO 2, 66 amu: COF 2 and its ionization product 47 amu: COF, all of them being predicted by the theory of FeF 3/PFPE reaction (3). | Document | AMS_2012.pdf | 165 |
152 Figure 6. Second run with associated mass spectrometry data.
The final experiment used a mixture of Fomblin Z25 and FeF 3 heated to 300°C using a 10 K/min ramp and cooled down to 40°C using a -10 K/min ramp, then maintained at 40°C for one hour. We could not see any measurable decrease once the temperature dropped below 260°C, which makes us think that no catalytic reaction takes place at ambient temperature, or that its rate does not allow one to see it with the experiment/equipment used and the time spent. The experiment has been made somewhat more difficult because of the absence of vacuum: gaseous compounds were noted to “contaminate” the test chamber, leading to sometimes false spectrometer readings. Moreover, some chemical species seem to be able to “stick” to the chamber from one test to another. This behavior has already been experienced within other test facilities such as the spiral orbit tribometer, even under a high vacuum. Given the complex spectra of the studied compounds, it was difficult to identify each species. Care should be taken during mass spectrometer analysis to know in advance what are the expected compounds and to have a reference spectrum for each of them, especially when their spectra are overlapping (which was the case here). The goal of the Pin-on-Disc experiments was to evaluate the influence of loading and temperature on the degradation mechanism. A total of 8 different experiments were conducted as shown in Table 1, using both temperature and mean Hertzian pressure as parameters. Figure presents the results of one particular test run at a 1200 MPa mean Hertzian pressure and at a 120 °C temperature. Notably, traces at 28 amu: CO and 47 amu: COF (not shown) were noted accordingly to the degradation mechanism theory (3). Interestingly, there was a continuous increase of these traces during the rotation of the tribometer. | Document | AMS_2012.pdf | 166 |
153 However, it has not been possible to compare the mass intensity of these traces between each experiment. Because of the presence of moisture inside the test chamber, the mass spectrum associated with water was exceeding the range of the mass spectrometer, thus rendering other readings inaccurate. Future tests should therefore carefully use amplification settings that do not make water traces exceed the range of the instrument. Table 1. List of conducted PoD experiments. Test 1 560 MPa, 20 °C Test 2 560 MPa, 80 °C Test 3 560 MPa, 120 °C Test 4 850 MPa, 20 °C Test 5 850 MPa, 120 °C Test 6 1200 MPa, 20 °C Test 7 1200 MPa, 120 °C Test 8 1900 MPa, 20 °C
Figure 7. Example of MS readings from a Pin-on-Disc test (red curve: 28 amu, blue curve: 47 amu).
0.00E+005.00E-051.00E-041.50E-042.00E-042.50E-043.00E-04 0.10.150.20.250.30.350.4 -1.00 1.00 3.00 5.00 7.00 9.00 11.00 intensityFriction coefficient duration [h]Friction coefficient 28 | Document | AMS_2012.pdf | 167 |
154 Conclusions
The current results have not allowed a precise reaction rate to be calculated at a given temperature, which would give a good estimation of this degradation mechanism’s activation energy and therefore allowed us to propose a safe range for operating this grease in space mechanisms. This knowledge would also have allowed us to estimate whether the degradation reaction is truly negligible at ambient temperature or not. In this case, even stopped stages should be considered harmful for the lubricant. However, we did manage to show that the presence of FeF 3 in the mechanisms definitely and considerably lowers the lubricant’s degradation temperature to a point that could be obtained through mechanical loading only (e.g., at ball/race micro contact level in a ball bearing). Looking at the gaseous compound release during the Pin-on-Disk tests, it is also important to take into consideration that the outgassing rate of the grease under loaded conditions significantly differs from an unloaded grease. A prediction of grease outgassing for a long lifetime application shall take these effects into consideration and adapt the quantity of grease accordingly. It has been noted that test facilities are easily contaminated by some chemical species. Care should be taken to follow proper “bakeout” procedures to ensure as much as possible that every trace of the previous test has been removed, especially in experiments where mass spectrometer data is important. Moreover, we would like to point out the importance of knowing in advance which spectra (and thus which peaks) are expected to be found in the mass spectrometer measurements, since expected species usually have complex spectra where various traces are overlapping. Further tests within a pin-on-disc tribometer and/or the spiral orbit tribometer will now focus on observing a real-case degradation to assess if it can truly be correlated to the degradation reaction we studied here. In particular, the influence of mechanical loading and temperature will be monitored. During these tests, care shall be taken to carefully choose mass spectrometer settings that allow proper comparison of the readings coming from the different experiments. We expect that these tests will help us to better define safe ranges of operation for the Z25 grease, and conclude on the autocatalytic effect existence or not. Bibliography
1. Carré, David J. Perfluoropolyalkylether Oil Degradation: Influence of FeF3 Formation on Steel Surfaces under Boundary Conditions. El Segundo : The Aerospace Corporation, 1985. 2. —. The Performance of Perfluoropolyalkylether Oils under Boundary Conditions. El Segundo : The Aerospace Corporation, 1991. 3. Marchetti, Mario. Aspects globaux et locaux de la mise en œuvre de la lubrification fluide en ambiance spatiale. Lyon : INSA, 2000. 4. Morales, Wilfredo and Mori, Shiguyeki. Reaction of Perfluoropolyalkylethers (PFPE) With 440C Stainless Steel in Vacuum Under Sliding Conditions at Room Temperature. Cleveland : NASA Lewis Research Center, 1989. 5. Pepper, Steve V. Effect of Test Environnment on Lifetime of Two Vacuum Lubricants Determined by Spiral Orbit Tribometry. Cleveland : NASA Glenn Research Center, 2011. 6. Zehe, M. J. and Faut, O. D. Acidic Attack of Perfluorinated Alkyl Ether Lubricant Molecules by Metal Oxide Surfaces. Cleveland : NASA Lewis Research Center, 1989. 7. Hsu, Stephen M. and Klaus, Erwin E. Estimation of the Molecular Junction Temperatures in Four-Ball Contacts by Chemical Reaction Rate Studies. ASLE Transactions. 1977, Vol. 21, 3. 8. Neave, J. H. and Joyce, B. A. The Origin of Spurious Peaks in Mass Spectra. Journal of Physics. Applied Physics, 1976, Vol. 9, 15. | Document | AMS_2012.pdf | 168 |
155 Trade Studies for a High Torque Density Planetary Gearbox
Jeffrey Mobley*
Abstract Sierra Nevada Corporation (SNC) has developed planetary gearbox design guidelines that significantly improve the torque capacity per unit volume for an optimized gearbox under a study funded through the NASA Phase II SBIR Lightweight Gearbox Technology Program. The design was based upon optimizing both the physical configuration and material selection of the gearbox design. As a part of this study, many different materials and material processes were evaluated for their suitability for use in lightweight planetary gearboxes for space applications. Calculation methods used to predict gearbox load ratings and expected life were also validated through test. This paper presents findings that have been extracted and edited from the final Phase II report published by SNC. Introduction Traditionally, planetary gearboxes used in aerospace mechanisms are based on conventional materials, processes, and designs. This usage has been primarily the result of rapidly paced schedules, and events taking place in a risk-adverse environment. Developing or honing new technology in the midst of a program is often impractical. This study provided the opportunity to research new combinations of materials and processes, as well as the opportunity to validate calculated gearbox ratings for capacity and life. The goal was to capitalize on this endeavor, in an effort to advance the technology used in the common gearboxes employed on planetary rovers and other similar space applications. Alternate materials and material processes have the potential to improve the load capacity and/or life of planetary gearbox designs. But many of those materials and material processes have little to no heritage in space mechanisms. The key is to develop optimum material combinations that balance the performance limiting factors within a planetary gearbox. One of the limiting factors is the radial load capacity of the planet bearing or bushing. Though ball bearing analysis remains uncomplicated, a general concern with journal bearings is the limited availability of data. Such data includes information regarding pressure, velocity, and pressure-velocity ratings for specific material combinations, and space applications which differ from standard journal bearing testing. Typical planetary gearbox designs used in space applications are not only based on the conservative guidelines set forth by the AGMA, but build additional margin throughout the system design. If confidence can be built in the calculated gear capacities for the applications and materials seen in space planetary gearbox designs, then it is possible to reduce the amount of margin in order to reduce the torque/mass ratio and still result in a highly reliable gearbox. Prototypical Gearbox Design A prototypical gearbox design was developed for use as both an analytical and test model. Every component in the prototypical gearbox design was carefully reviewed to optimize weight, facilitate the use of high capacity materials, and to ensure that manufacturability is maintained. The design was also easily reconfigurable with a variety of materials and finishes which facilitated its usage in further developing the proposed technologies. For the purposes of this study, the gearbox was designed to use a ring gear with five planetary gear stages with a ~1500:1 overall gear ratio, 48.3-mm (1.9-in) housing diameter (excluding flanges), and a 78.7 mm (3.1 in) overall length (excluding shaft extensions). Each stage uses optimized * Sierra Nevada Corporation, Durham, NC
Proceedings of the 41st Aerospace Mechanisms Symposium, Jet Propulsion Laboratory, May 16-18, 2012 | Document | AMS_2012.pdf | 169 |
156 geometry, materials and construction techniques that are best suited for the individual stage. The design goal was to achieve the maximum torque capacity with an output speed of 3.33 rpm. The overall gearbox weight was 0.59 kg (1.31 lb). An overall view of the resulting prototypical gearbox is shown in Figure 1.
Figure 1 - Prototypical gearbox design | Document | AMS_2012.pdf | 170 |
157 Prototypical Gearbox Rated Capacity The prototypical gearbox design was analyzed per ANSI/AGMA 2001-C95. Allowable bending stress versus cycles and allowable contact stress versus cycles was extrapolated from the AGMA standard based on the material hardness used. Stress factors were selected to match typical space planetary gearbox applications as much as possible. Based upon the analysis, the following capacities were predicted (assuming a 1.5 Reliability Factor for 99.99% reliability): 38.6 N-m (342 in-lb) operating torque against the endurance limit (1E7 Cycles) 56.2 N-m (497 in-lb) operating torque at 3.33 RPM for 205.1 hr (Pitting Life = Bending Life) 89.5 N-m (792 in-lb) for short term operation of 4 hr or less 137.5 N-m (1217 in-lb) operating torque for momentary operation (bushing dynamic limit, ≤5 minutes rated gear operation at this load) 205 N-m (1812 in-lb) static torque limit for gears (2.0 Safety Factor) Test Evaluation Parameters Bushing Materials For applications in which rolling element bearings are not suitable due to either capacity or packaging constraints, heritage planetary gearbox designs have traditionally utilized oil-impregnated SAE 841 bronze bushings. SAE 841 bushings have a Pressure-Velocity (PV) rating of 1.75 MPA-m/s (50 kpsi-ft/min) with a peak intermittent pressure of 28 MPa (4 kpsi). Often the bushing will be the limiting factor in the dynamic and/or static rating of the gear stage. An alternate material under consideration is Toughmet 3AT by Materion Brush, a spinodal/copper/nickel/tin alloy. Toughmet 3AT has a published PV rating of 4.6 to 9.0 MPa-m/s (132,000 to 260,000 psi-ft/min) depending upon the surface finish of mating parts. Maximum low-speed pressure was not provided and the testing that generated the PV ratings for the Toughmet 3AT material was performed at speeds of 1.5-2.0 m/s (300-400 ft/min) with additional lubricant added during the test. Space planetary gearboxes do not allow any lubrication to be added during life and often operate at speeds at or below 0.5 m/s (100 ft/min). It was decided to test both SAE 841 bronze and Toughmet 3AT bushings without providing additional lubrication and over a speed range closer to the typical space planetary applications. The purpose of the test was to determine if published PV ratings are valid and also to show a direct comparison between the two materials. Gear and Journal Surface Treatments The literature for Toughmet AT recommended Metalife Industries MLP as a surface treatment that could provide lower friction, higher PV ratings, and longer life in lubricant starved conditions. Metalife MLP is a thin dense chrome coating with a proprietary polymer compound added. Oerlikon recommended adding their Balinit C coating to gear teeth and mating surfaces of bushings to extend life and capability. Balinit C is an amorphous carbon-tungsten carbide coating (WC/C) with a high surface hardness and a low coefficient of friction that claims higher bearing load capacity, lower sliding wear, improved scuffing resistance, and reduced pitting particularly in applications with boundary lubrication conditions such as slow moving gears in contact. REM Chemicals recommended their Isotropic Superfinish (ISF) process which claims to eliminate breakin, reduce friction, and reduce contact fatigue. The ISF process involves a chemical conversion creating a micro thin soft layer that is removed by ceramic media. Multiple passes through the process resulted in removal of the peaks from the surface, while leaving the valleys as location for lubricant to remain during contact operation. | Document | AMS_2012.pdf | 171 |
158 Shot peening of gear tooth faces was also considered as a method to increase the localized surface hardness of the tooth, thereby increasing gear pitting fatigue resistance. Another expected benefit of shot peening of gears is an increase in the bending fatigue resistance at the root of the tooth. Many of these processes had not been previously performed on fine pitch gearing (64DP & 96DP) that is typical of lightweight gearboxes in space applications. These processes also had not been tested with space-grade lubricants which traditionally do not perform as well as their higher outgassing terrestrial counterparts. A development test was needed to determine which processes or combination of processes would add to the life and/or capacity of planetary gearboxes and may be suitable for future consideration. Capacity Validation AGMA analysis guidelines are typically conservative and do not always directly correlate to lightweight planetary gearboxes for space applications. Additionally, the parameters used in AGMA gear calculations were developed from testing larger gears, different environments, different materials, different lubricants, and different load conditions than typically experienced in Space applications. When testing is traditionally performed on a program, gearing is only tested to the loads and life specified for the application, and not tested up to or beyond the calculated rated loads and life of the design. Development testing was needed to validate the calculated allowable loads and life for the various base gear materials used and ensures that the calculations are not over stating the capability of the gearboxes. Testing Journal Bearing Test In order to directly compare combinations of bushing material (SAE 841 sintered bronze vs. Toughmet 3AT), shaft/planet finishes (as-machined vs. Superfinished vs. Metalife MLP vs. Balinit C), and lubricant (Bray vs. Pennzane), a journal bearing test was formulated. Various combinations were tested to determine relative pressure and velocity limitations for both high speed/low torque and low speed/high torque scenarios. Testing was performed in ambient temperature and pressure due to numerous required test setup changes to accommodate each combination. The oil impregnation from the SAE 841 bronze bushings coupled with relatively low hardness allowed the SAE 841 bushings to replenish the interface lubricant, absorb wear debris, and wear in any bushing surface damage to provide operation at levels higher than rated. The inherent hardness of Toughmet 3AT provided more uniform initial torque measurements, but in general these bushings failed at PV levels lower than the SAE 841 bushings and initial friction increases moved quickly to catastrophic increases in friction. REM Superfinishing displayed slightly reduced friction at the start of test, but overall results were mixed and there was not a conclusive improvement in life as a result of the process. Metalife MLP, when run against the Toughmet 3AT bushings, revealed a pressure limit of around 3.45 MPa (500 psi) at which point the coating began to break down regardless of speed. Balinit C shafts and gear blanks showed distinctly higher friction than other finishes and appeared to limit the SAE 841 material from recovering from initial onsets of increased friction. Pennzane formulated lubricants (Nye 2001 oil / 2000 grease) performed somewhat better than Bray formulated lubricants (815Z oil / 601 grease), but there was not a conclusive nor definitive trend. The optimal configuration was determined to be SAE 841 bronze bushings running on hardened steel shafts and planet bores with no additional surface coating or treatment except for surface finishes held as tightly as feasible. One of the lessons learned was that any chamfers, fillets, or edge breaks on the contacting surfaces should be performed prior to performing the final machining operation on the journal surfaces to minimize the risk of concentrated areas of high contact pressure. Heritage Phoenix SSI Gearhead Test In order to obtain advanced test results on some of the materials and finishes while the Prototypical Gearbox parts were being fabricated, spare gear parts from heritage Phoenix SSI Actuators were modified, assembled, and tested. One gearhead was assembled with Balinit C coated parts and one was | Document | AMS_2012.pdf | 172 |
159 fabricated with Metalife MLP coated parts. Both gearheads contained a mixture of shot peened and non shot peened parts. Both gearboxes were assembled with Toughmet 3AT bushings, Bray 815Z oil and Braycote 601 grease. The gearheads were operated at constant speed with increasing torque until a significant friction increase was detected at the input. Units were then disassembled and inspected. The Toughmet 3AT bushings displayed significant interface damage against both the planet posts and planet gears, in some cases seizing on the planet shaft. Due to the reduction in contact area and corresponding increase in contact pressure, shot peening increased the likelihood that both the Balinit C and Metalife MLP coatings would breakdown during operation. All of the Metalife MLP gear tooth wear surfaces showed some degree of coating breakdown, with the breakdown being more severe on the shot peened parts. Metalife MLP was determined to not be as durable as Balinit C for gear teeth under the pressures seen. Test results supported the journal bearing test data which indicated that Toughmet 3AT bushing material is not appropriate for this application. Additionally, it was determined that shot peening is a detrimental process for fine pitch gearing, particularly if surface coatings are desired. Prototypical Gearbox Test Gearboxes representing each finish (No Finish, Superfinish, Balinit C) and lubricant (Bray, Pennzane) combination were assembled for test. One combination, Bray/Balinit C, was not tested due to damage sustained to a carrier pinion during assembly. The remaining gearboxes were run-in at 14.1 N-m (125 in-lb) for 4 hours and characterized for baseline efficiency at loads up to 56.5 N-m (500 in-lb). Two gearboxes with No Finish were loaded with a static output torque ranging up to 282.5 N-m (2500 in-lb) to verify static capacity. Both units held the applied load with no sign of damage, verifying the static torque rating. One gearbox from each of the five configurations was tested through a 250 hr life test at 56.5 N-m (500 in-lb) external torque to verify rated life. This test was followed by 4 hours at 90.4 N-m (800 in-lb) to verify short term life. The Pennzane/No Finish, Bray/No Finish, and Pennzane/Superfinish gearboxes ran smoothly with no erratic torque spikes or significant changes in efficiency through the entire test. The Bray/Superfinish gearbox ran smoothly for the first 150 hours of the life test, and then started to show sporadic input torque spikes that would recover and recur, but never reached runaway levels. Efficiency testing on the Bray/Superfinish gearbox showed a reduction from 74% baseline to 56% post life. The Pennzane/Balinit C gearbox ran smoothly for the first 190 hours of the life test and then also started to show erratic torque spikes would recover and recur, but never reached runaway levels. Efficiency testing on the Pennzane/Balinit C gearbox showed a reduction from 73% baseline to 60% post life. One Bray/No Finish gearbox and one Pennzane/No Finish gearbox were operated up to and beyond the rated maximum momentary operational torque limit or 137.5 N-m (1217 in-lb) with no signs of erratic input torque which would have indicated bushing failure. The test was halted at the capacity of the input drive motor. The Bray gearbox reached 175.4 N-m (1552 in-lb) while the Pennzane gearbox reached 145.5 N-m (1288 in-lb) maximum momentary torque. Each of the life test gearboxes was disassembled and inspected. In general, Bray lubricant appeared darker and drier than Pennzane lubricant. The unfinished gearboxes revealed polishing at the gear meshes with no signs of significant surface degradation. The Superfinish/Pennzane gearbox appeared very similar to the No Finish/Pennzane gearbox. The Superfinish/Bray gearbox showed significant bushing wear in the 2 nd stage. The Balinit C gearbox with Pennzane (only Balinit C gearbox tested) showed significant bushing wear in the 2nd and 3rd gear stages with several of the bushings cracked. Even with significant bushing degradation, these units continued to function with reduced efficiency, but no sign of imminent catastrophic failure. The external static torque on one of the No Finish gearboxes was increased up to 425.7 N-m (3768 in-lb) with no sign of internal damage, significantly exceeding the rated capacity of the gearhead. The load applied was halted due to the output shaft rotating within the mating interface of the external coupling. | Document | AMS_2012.pdf | 173 |
160 Conclusion SAE 841 Bronze proved to perform at rated levels even in a lubrication-starved environment with space grade lubricants. SAE 841 Bronze also provides extensive additional margin between initial friction increase and catastrophic failure due to the ability to replenish lubricant and absorb debris. Toughmet 3AT did not perform to rated capacity most likely due to the inherent lubrication starved environment and the use of space grade lubricants instead of higher performing commercial alternatives. Unlike the SAE 841 bronze, initial onsets of friction quickly turned to catastrophic friction increases .
Metalife MLP provided lower initial friction than bare steel, but the coating broke down at relatively low operating levels, generating abrasive debris which resulted in a high wear rate. This coating may be suitable for bearing journals in high speed/low load applications but extreme caution would need to be taken. Metalife MLP did not appear suitable for use on gear teeth with any significant load. Balinit C displayed higher friction in journal bearings than bare steel and was detrimental to the performance of the SAE 841 bronze material. It is therefore not recommended for use against bushings. While Balinit C was more durable on gears than Metalife MLP, once the coating was broken down the debris was very abrasive and resulted in extreme wear. Balinit C may help fatigue life on higher speed, lower torque gear stages, but should be avoided in higher torque stages of fine pitch gears to prevent coating breakthrough. Shot peening resulted in significant surface roughness that could not be overcome by Superfinishing and resulted in swifter degradation of subsequent coatings of Metalife MLP or Balinit C. Shot peening is not recommended on 64DP or 96DP gears unless the process can be refined with extra fine media and reduced pressure, in which case additional testing should be performed. REM Isotropic Superfinishing provided mixed results. Initial friction levels were improved and there was less debris observed in the lubricant after gearbox run-in. However, overall gearbox life was not conclusively improved by the process. Superfinishing should be considered for 64DP and coarser gearing that cannot be cleaned and re-lubed after run-in. There remained concern over the impact of Superfinishing on the gear profile for 96DP gears (and finer). Gearbox rated capacities for short term, momentary torque, and static capacity were all validated through the testing of the standard No Finish gearboxes with additional margin shown. This data gives confidence moving forward that calculation techniques and assumed parameters are valid and gear designs can be pushed closer to their calculated limits to minimize mass and volume. Flight Application Sierra Nevada has incorporated design elements from this SBIR study into several programs that have been qualified, acceptance tested, and delivered. Significant programs included JPL Mars Science Laboratory Descent Braking Mechanism, JPL Mars Science Laboratory Low Torque Actuator Gearboxes (5 different gearbox designs), as well as other commercial Aerospace applications. References 1. “D21507 Final Report: Lightweight Gearbox Technology Program Phase II SBIR.” SNC 2. “ToughMet Plain Bearings – Performance with High PV.” Materion TechBrief 3. “Enhancing Toughmet 3 Bearing Performance with Metalife Hard Coated Steel.” Materion TechBrief 4. “Coated Components: Greater performance and reliability.” Oerlikon Balzers, 2010. 5. Winkelmann, Lane et al “The Effect of Superfinishing on Gear Micropitting.” Gear Technology March/April 2009 . | Document | AMS_2012.pdf | 174 |
161 Single Motion Actuated Shape Memory Alloy Coupling
A. O. Perez*, J. H. Newman** and M. Romano*
Abstract
The objective of the single-motion-actuated-shape-memory-alloy coupling (SMA2C) is to produce a small, single motion actuator used to secure and then release the solar panels on a CubeSat, or other very small satellite, upon command. The SMA 2C consists of a nickel titanium (NiTi) cylindrical shape memory alloy (SMA) press-fit into a stainless steel bushing, surrounded by a Kapton foil heater, a spring, a holding bolt, and a polyethermide isolation washer all in a structural housing. Heating the SMA above its activation temperature causes it to contract, permitting the spring to push the SMA, and therefore the holding bolt, out of the stainless steel bushing. This releases whatever the holding bolt is attached to from the CubeSat structure. Introduction
In the past, CubeSats, very small satellites of about one to four kilograms, have used various ways to secure their solar panels while in their CubeSat launchers. Some have been simply held closed by the walls of the launcher itself. The CubeSat’s solar panels would slide along the walls of the launcher and, when fully released, the solar panels would deploy. Sometimes the solar panel is secured by tying fishing wire around the solar panel and having a mechanism burn the wire for a controlled release. When this method is successful, it is quite simple and effective, but there have been premature releases occurring during vibration tests. In addition, there is a limit to the amount of force that the fishing wire can withstand, restricting its application to other types of deployments. SMA 2C is a release mechanism that addresses both control of deployment and capacity of holding force, which can be scaled for many applications. SMA 2C Background
NiTi is a shape memory alloy with particular properties for our application. Shape memory alloys are a class of materials that exhibit a mechanical change in property with a non-mechanical input such as a temperature change. SMAs undergo three types of transformations; Austenite, Twinned Martensite, and Detwined Martensite. In its natural state, SMA begins in its memory state also known as the Austenite phase. Once cooled it enters into the Twinned Martensite phase in which the crystal structures realign and produce a minimal change in overall expansion. If stress is not introduced during the Twinned Martensite phase and heated back up, the SMA’s crystal structure returns back to its Austenite phase and a minimal amount of overall contraction occurs. On the other hand, when stress ( ) is added in the cooled Twinned Martensite phase, an increase in strain ( expansion) of the Shape memory alloy brings it to the Detwinned Martensite phase wherein its crystal structure has been further modified. In order to go from the Detwinned Martensite phase back to Austenite phase, all that’s needed is an increase in temperature. This will cause the crystal structure to return to its original memory state (contraction). The amount of overall expansion is closely related to the amount of stress added during the Twinned Martensite phase. Figure 1 is a visual representation of this process. * Mechanical and Astronautical Engineering Department, Naval Postgraduate School, Monterey, CA ** Space Systems Academic Group, Naval Postgraduate School, Monterey, CA
Proceedings of the 41st Aerospace Mechanisms Symposium, Jet Propulsion Laboratory, May 16-18, 2012 | Document | AMS_2012.pdf | 175 |
162 Figure 1. Shape Memory Alloy phases [1]
SMA2C press fitting Assembly begins with a load cell in order to measure the amount of stress incurred on the NiTi during press fit. The outer bushing is placed inside the cradle and a little Isopropyl alcohol is dropped inside in order to clean and act as a lubricant during the press fit. The remaining cradle is then placed on top followed by the NiTi ring on the press pin. The press pin is then forced down causing the NiTi ring to be forced into the steel bushing as the load cell records the amount of force being asserted by the press pin. Figure 2 depicts the press fit assembly.
Figure 2. SMA2C Press fit [2] SMA2C holding force As mentioned above, the amount of stress induced between the Twinned and Detwinned Martinsite phase is directly correlated to the the amount of strain (expansion) that the NiTi undergoes. This | Document | AMS_2012.pdf | 176 |
163 expansion affects the amount of holding force when inside the bushing. In order to demonstrate this relationship, four different sample sets of NiTi rings were used with different outer diameters. They were each press fitted into a bushing of a certain diameter. The outcome was four sample sets, each with different interference measurements (difference between the outer radius of the NiTi ring and inner radius of the bushing). The sample with the greatest interference required greater force when press fitting the two pieces together, therefore increasing the stress induced. Because of this increased stress, the NiTi expands, leading to a greater force required to separate the two. Figure 3 plots the four different samples with varying interference and the amount of force needed to separate the NiTi ring from the bushing.
Figure 3. SMA2C holding force [2] SMA 2C assembly After the NiTi ring has been press fitted into the bushing, a polyethermide isolation washer is placed through the bushing in order to thermally isolate it from the remaining assembly and the CubeSat. The bushing is wrapped with a Kapton foil heater and this assembly is then inserted into the outer housing. A push plate and spring are inserted on the other end. This push plate will provide the little push required to separate when the NiTi is heated in a zero gravity environment. Figures 4 and 5 depict the final assembly. 00.511.522.5 012345Force (kN) Displacement (mm)Press-Fit Removal Forces for "G" Type SMA Rings Press-Fit Specimen #1 13 µm (0.0005” ) Press-Fit Specimen #4 76 µm (0.003” ) Press-Fit Specimen #3 51 µm (0.002” ) Press-Fit Specimen #2 25 µm (0.001” ) | Document | AMS_2012.pdf | 177 |
164 Figure 4. SMA2C final assembly [2]
Figure 5. SMA2C assembled [2]
SMA2C release The SMA2C housing would be secured to the bottom of the CubeSat. A retaining bolt would be inserted through the bushing and NiTi ring with the head of the bolt resting on the NiTi ring, and the other end secured to the solar panel. When commanded, current is sent to the Kapton foil heater, which then heats the SMA 2C micro coupling. Once heated, the NiTi will shrink allowing it to be released from the steel bushing. The push plate and spring would help nudge the NiTi ring along with the bolt out of the bushing.
Figure 6. SMA2C release [2] | Document | AMS_2012.pdf | 178 |
165 Conclusion
CubeSats, and other small satellites, require a mechanism to secure and to control the release of their solar panels and other deployables. One common practice is to use fishing wire to secure the deployable and cut it thermally at the desired time on-orbit to release it. Due to the limitation of the fishing wire’s strength and reliability, especially during vibration testing, other release methods may be useful. SMAs, when pressed fitted into another metal, provide a strong holding force which can be released with the addition of heat. The amount of holding force can be modified by simply adjusting the difference between the outer diameter of the SMA ring and the inner diameter of the bushing. Once heat is added to the bushing, the SMA contracts allowing the SMA and the bolt to release from the bushing. References
[1] D. C. Lagoudas – Editor, Department of Aerospace Engineering Texas A&M, Shape Memory Alloys Modeling and Engineering Applications ; Published by Springer Science + Business Media, LLC © 2008. [2] Crane, William Mike. Development of a Nano-Satellite Micro-Coupling Mechanism with Characterization of a Shape Memory Alloy Interference Joint. MS thesis. Naval Postgraduate School, Monterey, 2010. Print | Document | AMS_2012.pdf | 179 |
167 Development and Testing of a High Compact Stepper Motor Mechanism
Jörg Schmidt* and Greg Wright**
Abstract The Laboratory for Atmospheric and Space Physics (LASP) has developed for the Mars Atmosphere and Volatile EvolutioN (MAVEN) mission an Imaging Ultraviolet Spectrometer (IUVS) instrument. A grating flip mechanism (GFM), inside the IUVS instrument, is based on a highly compact and innovative stepper motor solution to drive an optic in two precise, bi-stable positions. Rotation angle between positions was 90°. This paper discusses major design restrictions, unique design characteristics and lessons learned from the development up through environmental and performance testing. Figure 1. IUVS mechanism overview * Phytron Elektronik GmbH, Grabenzell, Germany ** Laboratory for Atmospheric and Space Physics, University of Colorado, Boulder, CO
Proceedings of the 41st Aerospace Mechanisms Symposium, Jet Propulsion Laboratory, May 16-18, 2012 | Document | AMS_2012.pdf | 181 |
168 Introduction
Function of the GFM consists of turning a grating assembly inside two hard end stops. The position accuracy defines the systemic grating position. Very restrictive geometrical boundary conditions limited the degree of freedom for an actuator solution – especially with respect to axial length and diameter for the actuator. Standard components or “of the shelf” products were not applicable. Low weight, high rigidity and functional safety were fundamental design criteria. The long-term cleanliness of the grating represents a systemically size for the spectroscope. To keep the probability and the various ways of grating surface contamination in a very low level, an effective design strategy had to meet all these criteria. This risk mitigation was the key point in the complete design concept. A much more compact construction had to be developed. Mounted on the cover was the grating assembly with the counter balance mass. The rotation angle was 90°. Two hard end stops on both sides limit the rotation angle. Electrical limit switches detect both end positions. To facilitate the bi-stable positioning we added a flexible coupling between cover and planetary gear shaft. Our purpose was to turn the grating assembly very smoothly in to the end positions. A torque preload should hold the grating in the end position under unenergized conditions via detent torque. Motorization was done by a special modified hybrid stepper motor size VSS 25.200.0,1-X from Phytron. Figure 2. Integrated stepper motor mechanism planetary gear cover flexible coupling double row hybrid bearing (instrument main bearing) stepper motor central housing | Document | AMS_2012.pdf | 182 |
169 Main Functional Requirements
Thermal Constraints The operating temperature range from -10°C (safety range -30°C) up to +40°C creates some different friction characteristics. Each lubricated bearing needs a little higher torque under cold conditions. At begin of the project it was not possible to calculate the friction characteristic for the double row instrument bearing. Equivalent data were not available from the manufacturer. Due to the special construction, quality criteria, preload and specified materials, no similar bearing was usable for a representative friction test at that time. 36 balls in each bearing race, 72 balls in total. 144 wet lubricated single contact points creates a higher friction uncertainty. Torque loss with affects to the functional safety was a critical issue. Agglomeration effects during long time storage were an additional uncertainty. Other bearings in the mechanism and the lubrication inside the planetary gear increase the complex of friction problems. The thermal characteristic influences all parameters. These are well known effects. On the other hand, the stepper motor, driven by a constant voltage driver, must create enough torque to overcome all static and dynamical friction parameters. The integrated 2-stage planetary gear with a ratio of 49:1 increases the available stepper motor torque. Friction Torque Separation To minimize the thermal induced torque reduction and uncertainty we try to split the consequences over the time. For a better understanding we must describe the process, to drive the grating assembly in to a hard end stop, in more detail. The stepper motor with 200 steps per revolution (=1.8°/full step) operates normally with a small magnetically load angle, depending on the torque load. The planetary gear have a backlash of 35 arcmin and the very exact calculated and qualified flexible stiffness of the coupling allows us to drive with 4 motor full steps in the hard end stop. So, if we turn out of the end stop we can use the: - stepper motor magnetic load angle - planetary gear backlash - flexible coupling stiffness for a friction torque separation. If we turn the grating assembly out of the torque preloaded end position, the motor must not overcome all friction parameters at same time. In the first milliseconds, the complete motor torque is only necessary to overcome the motor internal friction torque (bearings, detent torque). Motor shaft began to turn. The planetary gear is now positive preloaded due to the flexible coupling and began to turn with a much lower friction torque as normally required. The backlash reduces the coupling preload slowly and absorbed some vibrations (function like a damper). Some milliseconds later the planetary gear began to turn. In a third step on the time line, motor and planetary gears were turning; the flexible coupling was used as a torque peak damper to overcome the friction torque of the instrument bearing. At -10°C, we need 34 mA and at -30°C only 37 mA starting current. So the engineering teams from LASP and Phytron played with the characteristic and performance of different design elements to create friction risk mitigation. To drive in a hard end stop is from a friction and torque point of view not so critical due to the lower dynamical friction values. The worst case was to drive out of a hard end stop. | Document | AMS_2012.pdf | 183 |
170 Final Concept Approach
Flexible Titanium Coupling To get enough safety margins in the flexible coupling design, following the wisdom “expect the unexpected”, we start with some different designs. Figure 3. Evolution steps flexible coupling Nearly 30 different models were calculated to find an optimal flexure structure. The stiffness was defined and the dimensions very hard restricted. Material stress reduction and the radial and axial dimensional changes were limited, too. We could reduce the internal stress from 724 MPa down to 54 MPa (= ~8%). Both parties recalculated the final FEM model. LASP with ANSYS, Phytron used COSMOS. The results correlated in the stress and dimensional changes very exact. To reduce any risk in the later assembling and test phase, we decided to produce three different flexure structures in the couplings. So, three versions were manufactured and tested respective to the torsional stiffness. Calculated and final measured stiffness values were different. Similar results were found in [1]. Figure 4. Structure of flexible couplings Intensive studies of comparable flexible structures, which were used in other mechanisms e.g., titanium butterfly pivots [2] and their final tested characteristics, creates a higher confidence level for our design approach. | Document | AMS_2012.pdf | 184 |
171 Central Housing Structure The only constructive way to be able to press many requirements in a small space was to use a well coordinated conceptual design in hybrid joining technology. The central housing used all the advantages of the accuracy of CNC machines. At the same time, it represents the basic structure without soft joints. To achieve excellent out gassing properties, the cover envelops the entire internal structure on 3 sides. Desorbing products will be dissipated controlled in one direction. Figure 5. Main structural titanium parts and complete mechanism Final Grating Flip Stepper Motor To obtain detailed performance data from each assembly phase, several tiered test sequences have been defined. The performance data were therefore known at any time. The irreversibility of some assembly processes requires a precise line of action with parts which must absolutely comply with the requirements. Figure 6. Final tested motor assembly | Document | AMS_2012.pdf | 185 |
172 The complete unit can be connected with the IUVS instrument structure via 10 internal threads. Heli coil inserts prevent a fretting between the materials. The cover contents similar threads for the grating flip mechanism. Both mounting points are very close together. The stiffness of the large instrument hybrid ball bearing dominates the overall stiffness. Figure 7. Final tested motor assembly The flexible coupling separates two mechanical oscillating systems. One system is defined by the stepper motor, the planetary gear, and the upper part of the coupling. Vibration source is represented by the stepper motor. Second system is defined with inertias of the grating flip mechanism, the instrument ball bearing and the cover. Static and dynamic friction in side ball bearings and g ears can have far-reaching repercussions on the damping properties.
Main Test Results Outgassing Test Results A residual gas analyzer (RGA) was used to measure the volatile impurities in the assembled actuator at +40°C. Results of the test indicate the actuator is suitable for the strict requirements of a FUV optical instrument – i.e. virtually indistinguishable from a clean vacuum chamber. Vibration Testing The EQU actuator was assembled in to the GFM life test unit (LTU) and subjected to the corresponding qual-level vibration environment. Test levels included sine burst to 17 Gs and random vibe to over 10 Gs rms per axis. Following the testing the actuator performed smoothly and passed all post-test inspections. Life Testing Following the vibration testing the LTU was subjected to a thermally controlled vacuum life test which included over 6,000 cycles at a range of temperatures. The actuator met all operational criteria from -30° C to +40°C and passed the life test successfully. Optical Repeatability Testing The LTU was driven into the hard stop over 200 times, and the optical orientation of the grating was measured each time using an autocollimator. The position repeatability was measured at less than 10 arc-sec, which meets the mechanism requirements. | Document | AMS_2012.pdf | 186 |
173 Conclusions and Recommendations
1. A sequence of several manufacturing processes reduces the coupling stiffness in a range of approx 40%. Result of different material properties and the unpredictable behavior in the rim zone. Our strategy, to create 3 different coupling versions, was very helpful at the end. Real tests were highly recommended in an early project phase.
2. Trust-based cooperation between the design teams at a very early stage of the project is the basis of innovative products. 3. The distribution of the entire starting friction in several temporally separated events represents an efficient way. 4. Adjusting rings made of titanium with fine threads creates difficult problems during assembling. Very precise inspection and deburring under the microscope is necessary.
5. The characteristic of both mechanical oscillating systems must match exactly together. Meant with “match” is naturally “out off-tune”. At worst case, the system creates resonance or bouncing effects. In a complete representative "plug and play" system test, the performance characteristics should be verified in an early project phase. Alternative settings shall be provided in the degrees of freedom of control electronics or software intelligence.
Acknowledgement The authors would like to thank Steve Steg and Heather Buck from LASP for some technical concepts, FEM calculations and technical discussions, Wahid Lahmadi from Phytron Inc. for all the project management support on this project. References
1. Santos, I., et al., "HIGH ACCURACY FLEXURAL HINGE DEVELOPMENT", ESMATS 2005 2. Henein, Simons, et al., "FLEXURE PIVOT FOR AEROSPACE MECHANISMS", ESMATS 2003 | Document | AMS_2012.pdf | 187 |
175 Cryogenic Temperature Testing of NEA Fuse Wire Mechanism
Edwin Vega and Geoff Kaczynski
Abstract
NEA has over 10 years of continuous innovation in the field of space rated non-pyrotechnic release mechanisms. At the heart of our technology is the NEA patented Fuse Wire Assembly (FWA). NEA utilizes our patented FWA design across all our product lines to initiate the release operation of our restraint and release devices and our battery bypass switches. Most recently, as part of a deep space mission, cryogenic temperature testing was performed on our FWA down at 18° Kelvin (K). The testing discussed in this paper validates that the FWA, which is a critical component of the overall mechanism design, is robust enough to reliably operate in space at extremely low power levels and temperatures. In this paper, NEA will describe the cryogenic temperature tests along with test results. As the space community explores further into our solar system and beyond, this validation testing of our product enables new found confidence that existing technologies can be used for fielding deployable structure in deep space. Introduction All major space agencies and organizations are funding missions for spacecrafts to explore our solar system and beyond. As soon as a spacecraft leaves the earth orbiting environment, it experiences extreme cold temperatures with the background radiation of the universe. Today’s missions require several staged deployments during multiple phases of their missions, such as entering orbit, decent, landing and deployment of instruments. For today’s missions, the restraint and release device must be proven to operate in deep space cryogenic temperatures to ensure mission success. The NEA release mechanism is used to provide restraint and release functions for critical deployment operations on spacecraft such as solar arrays, antennas, radiators and payloads after launch. The FWA in the release mechanism is a critical element. Each release mechanism’s FWA is single or redundantly initiated. Actuation time variation between multiple mechanisms FWA is less than 10 milliseconds when a nominal firing pulse is applied to all units simultaneously.
Basic Design
The patented NEA FWA design is simplistic in operation and design. Upon receiving a specified electrical pulse of 1.2 amps or greater, the fuse wire breaks starting the release sequence of the preloaded release rod. The shock induced by release of the tensile load is typically less than 300 G’s. Each FWA is fully refurbished after each ground-based test actuation. The principle of operation is as follows: A fuse wire is wrapped between two electric terminals. A process controlled load is placed on the fuse wire. When current is applied via the spacecraft firing system, the fuse wire heats lowering its yield strength. The fuse wire yields once the applied load exceeds the strength of the fuse wire. Breaking of the fuse wire allows the release system to activate and allow deployment. * NEA Electronics Inc., Moorpark, CA
Proceedings of the 41st Aerospace Mechanisms Symposium, Jet Propulsion Laboratory, May 16-18, 2012 | Document | AMS_2012.pdf | 189 |
176 The electrical characteristics for the fuse wire assembly are time and current dependant as shown in Figure 1.
Figure 1: NEA Mechanism Actuation Curve Extensive test data has shown that time for release is not significantly influenced by external operating temperature. See Figure 2 for data. 020406080100120140160180200 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10Actutation Time (Milliseconds) Actuation Current (Amps) | Document | AMS_2012.pdf | 190 |
177
Figure 2: Fuse Wire Actuation Time vs. Temperature Curve
Cryogenic Temperature Testing
The intent of the Cryogenic Temperature testing was to verify if a typical NEA release mechanism could withstand cryogenic temperatures and successfully actuate to release the preload at these temperatures. Two units were used for the Cryogenic Temperature Test. These two units were flight worthy NEA 9103 release mechanisms. A preload of 11,120 N (2,500 pounds) was applied as shown in Figure 4.
Figure 1: Cryogenic Temperature Benchtop Setup for the NEA Mechanisms A voltage of 1.5 Vdc from a D-size battery was set across the switch plates. A 0.13-mm (0.005-in) thick stainless steel leaf spring was included as a temperature compensation element. When the mechanism was actuated, the ejection spring forced the insulator and corresponding switch plate away and caused | Document | AMS_2012.pdf | 191 |
178 an open circuit. This verified successful release of the preload. The noise from the hardware hitting the bottom of the chamber also served as evidence of successful release of the preload. The fuse wire resistances were taken at lab ambient temperature.
Figure 2: Cryogenic Temperature Test Setup for the NEA Mechanisms Due to a limited number of chamber output leads, only the redundant circuit was actuated. The redundant circuits were wired so that each shared a common ground. The primary circuits were grounded to the plate so as to assist in reaching the temperature. Thermal diodes were taped to the cover of each unit. Once the test fixture was fastened to the mounting head of the cryogenic chamber, the resistances were measured again at lab ambient temperature. The chamber as shown in Figure 5 was set to vacuum conditions and the temperature was decreased using liquid helium as the cooling media. The chamber achieved temperature stabilization at the targeted temperature of 10K after approximately 20 hours. After temperature stabilization was met, the unit temperature for both mechanisms was approximately 17K. The temperature of the mounting plate was 10.2K. Fuse wire circuit resistances were taken once again. The continuity was checked for each of the switch plate circuits. The switch plate circuit for the first mechanism displayed continuity. The switch plate circuit for the second mechanism reads open. It was determined that the second unit would be actuated first, due to the fact that the switch plate circuit was open. Once the second unit was actuated, the switch plate circuit was re-connected. The switch plate pulse dropped from approximately 0.5 V to zero, indicating successful separation. The release rod and associated hardware could be heard hitting the bottom of the chamber. The unit successfully released the preload. The temperature of the unit dropped immediately after actuation. | Document | AMS_2012.pdf | 192 |
179 The first unit was then actuated following the same procedure. Once the unit was actuated, the switch plate circuit voltage dropped from approximately 0.5 V to zero, once again indicating successful separation. The release rod and associated hardware could be heard hitting the bottom of the chamber. The unit successfully released. The pulse for the trigger circuit was shown re-connecting. This was an anomaly in the test setup. The captured scope plots of the actuation events are shown in Figure 6 and Figure 7.
Figure 3: Actuation Plot for the First Release Mechanism
Figure 4: Actuation Plot for the Second Release Mechanism | Document | AMS_2012.pdf | 193 |
180 Conclusions
Cyrogenic testing has shown that the NEA FWA and mechanism reliably actuate at cryogenic temperatures. The fuse wire circuit resistance for each of the devices decreased by no greater than 0.3 ohms from lab ambient conditions compared to cryogenic vacuum conditions. The mechanisms successfully released at approximately 16K in both loaded (2,500 lb) and the unloaded conditions. Post test inspections revealed no physical degradation due to actuation. In particular, the restraining wire remained intact which would lead to a typical dissipation of strain energy and therefore a low shock output. | Document | AMS_2012.pdf | 194 |
181 Developmental Testing of Electric Thr ust Vector Control Systems for Manned Launch Vehicle Applications
Lisa B. Bates* and David T. Young**
Abstract
This paper describes recent developmental testing to verify the integration of a developmental electromechanical actuator (EMA) with high rate lithium ion batteries and a cross platform extensible controller. Testing was performed at the Thrust Vector Control Research, Development and Qualification Laboratory at the NASA George C. Marshall Space Flight Center. Electric Thrust Vector Control (ETVC) systems like the EMA may significantly reduce recurring launch costs and complexity compared to heritage systems. Electric actuator mechanisms and control requirements across dissimilar platforms are also discussed with a focus on the similarities leveraged and differences overcome by the cross platform extensible common controller architecture. Introduction
The potential for ETVC systems to significantly reduce recurring launch costs, complexity, weight and volume, compared to electro-hydraulic systems of equivalent performance and reliability, soon may be realized on large launch vehicles for human space flight. ETVC systems have been used in the Apollo and Space Shuttle programs in the past. But conditions unique to the launch environment have up to now restricted their use for manned spaceflight to less powerful in-space applications. The lack of a suitable electrical power source and approved human rated power electronics that could be qualified to the launch environment, as well as the susceptibility of high voltage electrical power systems to corona discharge, have placed severe limitations on the power of these early manned systems. Renewed interest in ETVC systems for high power launch vehicle applications is due to advances in key enabling technologies related to the source and control of electrical power. High rate lithium ion batteries, high-voltage, high-current insulated gate bipolar transistors (IGBT) and Field Programmable Gate Arrays (FPGA) are among the maturing technologies incorporated into the cross platform extensible controller architecture of the ETVC system tested and described in this report. The controller, battery modules and the integrated ETVC system based on a developmental EMA are the result of an internal research and development effort by Alliant Tech Systems, Aerospace Systems Group and Moog Inc., Space and Defense Group. Testing was performed in cooperation with NASA George C. Marshall Space Flight Center at its Thrust Vector Control Research, Development and Qualification Laboratory. Among these maturing technologies, high power switching electronics such as the IGBT in particular has made it possible to further simplify actuator mechanisms and eliminate certain mechanical failure modes. IGBTs have been used extensively in the electric vehicle industry, not only to create the inverter circuits needed to power 3-phase Brushless Direct Current (BLDC) motors, but also as a key component in regenerative braking circuitry. How these technologies can reduce the mechanical complexity of an ETVC actuator will be seen in a comparison of the developmental EMA actuator mechanism with that of Apollo and Space Shuttle EMA. The work presented in this report represents one phase in an ongoing development program aimed at demonstrating the maturity of high power ETVC systems and components for manned launch vehicle * NASA George C. Marshall Space Flight Center, Huntsville, AL ** Raytheon – Jacobs ESTS Group, George C. Marshall Space Flight Center, Huntsville, AL
Proceedings of the 41st Aerospace Mechanisms Symposium, Jet Propulsion Laboratory, May 16-18, 2012 | Document | AMS_2012.pdf | 195 |
182 applications. The next phase of this effort will be to update a multi-channel electro-hydrostatic actuator (EHA) by adapting it to the common controller architecture while again utilizing advanced lithium ion batteries as its power source. In preparation, testing of the EHA in its present form was also carried out to baseline its performance. Objectives of the Experiment The primary goals of this experiment were to verify integration of the ETVC developmental hardware and to demonstrate functionality of the complete system. Flight specific performance requirements were not set. Moreover, this test was carried out with fewer battery modules than would be needed to achieve the full power capability of the actuator. As such, the system under test was considered to be underpowered and it was necessary to design control parameters accordingly so that peak power demands under the applied load would not exceed the capabilities of the available battery modules. Performance measurements are, therefore, meant more to indicate general functionality of a representative class of Thrust Vector Control (TVC) system, rather than in meeting a particular vehicle requirement. Key test objectives: • Functional integration of ETVC components • Controller parameters tuned “in the field” • Peak power draw and voltage droop controlled to acceptable levels • Step and frequency response as expected for underpowered performance • Battery cell temperatures stay near ambient • Repeated operation on a single battery charge For the functional demonstration, dynamic loads due to inertia and spring forces were supplied by two large Inertial Load Simulators located at the Marshall Thrust Vector Control Research and Development and Qualification Laboratory.
Figure 1. Electromechanical (left) and Electro-Hydrostatic (right) Actuators
In addition to the dual channel developmental EMA shown in Figure 1, tests were performed with the four-channel EHA also shown, to baseline its performance. This report presents data and the results of analysis for only the EMA. | Document | AMS_2012.pdf | 196 |
183 Background
Control of a launch vehicle during ascent implies the ability to direct the vector of the thrust that it produces. Typically a pair of linear actuators, positioned so as to rotate an engine or nozzle about its bearing along orthogonal planes, act together to define a resultant thrust vector. High inertial load and the requirement to operate from sea level to near orbital altitude are characteristics of launch vehicle applications. This is in contrast to the low load, vacuum conditions of in-space applications. High powered TVC systems are needed to react against the inertia of an engine or nozzle, as well as against the stiffness of an engine gimbal or flex bearing, vehicle structure and propellant flex lines, at the slew rates necessary to maintain stable control of the vehicle throughout all phases of flight. Heritage Electro-Hydraulic Actuation Historically, high power demands could be met only by hydraulic systems. One means by which these systems could derive enormous amounts of hydraulic power was by accessing a pressurized propellant line in a liquid fueled rocket engine, such as in kerosene-based engines, at the cost of a slight performance loss to the engine. However, not all rocket propulsion systems are compatible with this approach. Therefore, auxiliary power generated by hydraulic turbo pumps and dedicated propellant systems have also been used in heritage hydraulic systems.
Figure 2. Heritage TVC System
The complex arrangement of discrete hydraulic components, seen in Figure 2 for the Space Shuttle Solid Rocket Booster, was typical of high power heritage TVC systems. It was common for such systems to have a long and elaborate process flow associated with their assembly. A major detractor that opponents of this particular approach often cite is that in order to drive the turbo pumps that generate the needed hydraulic power, these systems typically relied on the decomposition of toxic monopropellants such as hydrazine, a known carcinogen with costly storage, handling and safety concerns. To deal with concerns about toxic monopropellants one could simply replace the heritage turbo pump with a fixed speed electric motor driven variable displacement pump, while leaving the rest of the heritage system unchanged. The trade in this case would likely be a slight increase in system mass and volume due to the lower energy and power densities of electrical power sources compared to that of monopropellant powered hydraulic turbo pumps. Another way to deal with these concerns would be to retain the turbo pump but substitute a less toxic monopropellant with equal performance, such as an ammonium dinitramide based liquid monopropellant. However, in either case, leaving the rest of the | Document | AMS_2012.pdf | 197 |
184 heritage system unchanged means that the complex assembly of the discrete components of the overall system still remains a labor intensive process. While there are certain benefits that may favor heritage electro-hydraulic systems in a trade of alternate TVC approaches, these are likely to be only in the short term. As available stores of heritage hardware dwindle and mature enabling technologies continue to further advance high power ETVC capabilities, these perceived benefits will inevitably diminish. ETVC Systems There are a variety of possible ETVC systems available based on the type of actuator and source of electrical power. But, in general, their inherent simplicity, compared to heritage systems, means that they can be expected to have lower operating costs because of simpler, less hazardous ground operations. The potential for less overall system weight and volume is also a possibility, in spite of the fact that electrical sources are not as power dense because of the many discrete heritage system components that can be eliminated. For example, a complete ETVC system like the EMA system tested and described in this report, or a similar one based on the mentioned EHA, would consist of only a pair of actuators along with a set of controller boxes and a bank of battery modules such as those shown in Figure 3.
Figure 3. Lithium Ion Battery Module (left) and Controller The simplicity of the EMA and EHA systems and the streamlining of ground operations that they afford is a feature shared by another type of ETVC system, the Integrated Actuator Package (IAP). The best way to think of an IAP is as an entire electro-hydraulic system self contained within each actuator. The IAP takes advantage of the approach mentioned earlier of using a fixed speed electric motor and a variable displacement pump to generate hydraulic power. This strategy, partly because it does not have to deal with propellants or turbine exhaust, allows the IAP to more readily integrate each of the components of a conventional electro-hydraulic system into a self-contained package. Like the heritage electro-hydraulic system, servo valves are used to continuously interpret low power electric command current, measured in milliamps, and, thus, regulate hydraulic power to either side of the actuator main piston. In this way, the IAP is very similar to the EHA. Electro-Hydrostatic Actuation The electro-hydrostatic actuator is another approach available for ETVC systems that possesses both electrical and hydraulic power system attributes. Unlike purely hydraulic TVC systems, EHA systems rely on an electric motor driven positive displacement pump, incorporated into the body of the actuator, to generate hydraulic power and meter fluid to and from a hydraulic cylinder. Like the IAP, all hydraulic components including the fluid reservoir, manifold, filters, etc. are incorporated into the body of the actuator, simplifying system integration and reducing the total volume of hydraulic fluid used in the overall system. The difference between these self contained systems is in the type of pump used and in the way position commands are interpreted, which for the IAP involves servo valves. | Document | AMS_2012.pdf | 198 |
185 Figure 4. EHA System Elements
The schematic in Figure 4 shows the system elements of a typical EHA system and can be used to understand its operation. Position commands interpreted by a controller (not shown) are used to continuously update speed and direction of a reversible, variable speed, 3-phase BLDC motor. This motor drives a fixed displacement pump creating a hydrostatic pressure difference across the hydraulic cylinder. The fixed area of the cylinder piston translates this pressure difference into a force proportional to the speed of the motor which acts against the applied load. Finally, position and velocity feedback (not shown) is used by the controller to close the control loop on the commanded position. The EHA, which was also tested, is a four channel version of this same arrangement. Each of four identical channels responds independently to what are nominally the same commands. The hydrostatic pressure developed by all four channels combines at the hydraulic cylinder to create a net total pressure and sizing is sufficient to tolerate the failure of two channels without loss of the targeted performance. The blocking valve is used to remove an errant channel from the system by equalizing its pressure contribution and pressure relief valves are provided for safety. Sizing of the reservoir, depending on the application, is either determined by peak power demands or is based on the thermal capacity of the total hydraulic volume and the mission duration. Electromechanical Actuation The EMA, the simplest of all ETVC actuator types, is nothing more than a mechanism that converts electrical energy into the torque of a rotating variable speed motor and then into linear motion through a mechanical transmission. This ultimately puts energy into the motion of an engine or nozzle mass that the EMA must also be able to absorb as it brings this motion to a stop. In the case of the EMA tested, a 3 phase BLDC motor is used to both add and remove kinetic energy to the overall system. Many electric automobiles use BLDC motors and high power solid state switches to recover energy and improve mileage. But, whether recovered by the electrical power source, or simply dissipated through a resistive load, high power solid state switches such as the IGBT make it possible to handle high levels of excess kinetic energy electronically through motor torque. As a result, the mechanical brakes and clutches of older systems, with their potential for failure due to contamination and wear, are no longer necessary. The architecture of the EMA system that was tested allows it to be single fault tolerant, with the redundancy of two active channels. Identical position commands, which on a vehicle would come from a flight computer, are sent to a motor controller on both of its command channels and telemetry feedback from sensors within the actuator is returned. Position feedback is measured directly by a dual channel linear variable differential transformer (LVDT). On the other hand, actuator velocity, the rate of change of the actuator rod end position, is not always a direct measurement. Often it is derived from electric motor velocity, which is much faster prior to gear reduction and, therefore, offers greater resolution. Motor velocity can be monitored using a generator or similar such device. Motor position can also be acquired | Document | AMS_2012.pdf | 199 |
186 using a resolver or an encoder. In the case of the developmental EMA, resolvers are used for both motor position and motor velocity. As illustrated in Figure 5, the outputs of two identical 3-phase BLDC motors are combined through spur gears to create a torque-summed moment upon a common ball screw mechanism. The ball screw is used in a rotating nut / translating screw configuration. It should be noted that because two channels are combined at a common ball screw, mechanical redundancy is lost at this point. There are several potential failure modes that can be identified for the common ball screw. Seizure of the ball screw mechanism preventing motion of the actuator at an inopportune time can be catastrophic. For the EHA and IAP, hydraulic power transmission through a main piston cylinder virtually eliminates potential jamming concerns assumed by electromechanical transmissions. The strategy for dealing with such failure modes in the EMA that was tested is to size each motor and drive train sufficiently so as to provide adequate torque to overcome some degree of potential mechanism seizure. It has been asserted by the manufacturer that deformation of the ball or race, by this means, or through wear, as well as any potential hazard of contamination, would likely result in only degraded actuator performance, but not failure. Based on manufacturer studies, it is expected that under such degraded conditions, the ball screw would simply behave like a nominal Acme screw mechanism.
Figure 5. Dual Channel developmental EMA and Controller Precedence for a multichannel EMA actuator with common drive train components can be found in the first ever ETVC system developed for manned space flight. The Apollo Service Propulsion System (SPS) which was relied upon to perform its mission critical trans-lunar injection maneuver, successfully employed a mechanically similar approach for its EMA. However, the Apollo EMA differed in its redundancy scheme in that it employed an active / standby system. The Apollo SPS was located within the Service Module of the Apollo spacecraft and utilized only after the Saturn launch vehicle had carried it into orbit. Therefore, the Apollo EMA is considered to be an in-space propulsion system, and this fact likely influenced the choice of an active / standby approach. Another example where ETVC have been effectively used in manned space flight is found in the Space Shuttle Orbital Maneuvering System (OMS). The OMS, located near the aft end of the Space Shuttle Orbiter provided thrust to perform orbit insertion, orbit circularization, orbit transfer, rendezvous and de-orbit. These were maneuvers performed after Space Shuttle Main Engine cutoff, and, for this reason, the OMS is also considered to be an in-space propulsion system. | Document | AMS_2012.pdf | 200 |
187 The Apollo and Space Shuttle EMA were both two channel actuators, which nominally operated one active channel, the primary channel, while the other secondary channel was reserved as a standby in the event of primary channel failure. This differs from the developmental EMA, which nominally operates two active channels, a scheme that is considered to be more reliable because of the probability associated with the standby channel not being available when called upon, a risk more often accepted for in-space propulsion than for launch because it is not always practical to require in-space systems to be active prior to launch. Figure 6. Active / Standby EMA Mechanisms Careful consideration of the two mechanisms illustrated in Figure 6 will provide insight into some of the ways in which redundancy can be achieved by EMA systems and of their relative merits. The mechanism on the left utilizes two clutches with each motor to control the direction motor torque is applied to a common jack screw through a common drive gear. Braking is achieved by applying torque opposite the direction of travel. The nut tube and the jack screw on which it translates as the jack screw turns are single point of failure components. Likewise, the common drive gear and associated shaft and bearing are single point of failure components because these components are also shared by both channels. The mechanism on the right of Figure 6 illustrates how redundancy can be achieved throughout the entire drive train including translational motion. As seen in this illustration, the nut tube will translate along its external spline due to the rotation of the jack screw or due to the rotation of the nut tube as driven by the rotation of the internal spline bushing. Disk brakes are used in this mechanism to prevent movement of the secondary standby channel while the primary channel is active, or vice versa. Significantly, these brakes are also used to arrest movement of the active channel and, thus, dissipate energy as motion of the engine nozzle is stopped. Power and Control Considerations Technology innovation efforts such as the More Electric Aircraft initiative by the Air Force and efforts by the commercial aircraft and automotive industries have led to many new advances in power electronics and direct current power sources. For aircraft applications, a bus voltage of 270 volts has emerged as more or less an industry standard. But, launch vehicle applications have unique considerations because of the altitudes traversed for which this voltage may be a concern. The dielectric constant of the air changes with altitude, making it easier at higher altitudes for high voltages to break down the dielectric barrier of the air in a process known as corona discharge. At still higher altitudes, susceptibility to corona discharge vanishes as the vehicle enters near vacuum conditions, which is why high voltages do not exhibit this particular concern on orbit. But, for launch vehicles this phenomenon can be potentially disruptive to sensitive electronics elsewhere on the vehicle, as well as to the TVC system itself. At sea level, thousands of volts are required to induce the onset of a corona discharge. For aircraft, a bus voltage of 270 volts is not a corona concern because it is well below the corona onset voltage at the altitudes where most aircraft fly. Many factors influence a corona discharge event. Atmospheric pressure and constituency, bus voltage and frequency, as well as conductor geometry, are all factors that play a | Document | AMS_2012.pdf | 201 |