Development of a Highly Sensitive, Highly Stable Micro-Newton Thrust Balance

Transcription

1 Development of a Highly Sensitive, Highly Stable Micro-Newton Thrust Balance IEPC Presented at the 35th International Electric Propulsion Conference Georgia Institute of Technology Atlanta, Georgia USA Franz Georg Hey and Max Vaupel Airbus, Friedrichshafen, Germany Claus Braxmaier DLR, Institute of Space Systems, Bremen, Germany Universität Bremen, ZARM, Bremen, Germany Martin Tajmar Technische Universität Dresden, Institute of Aerospace Engineering, Dresden, Germany and Alexander Sell, Karlheinz Eckert, Dennis Weise, Noah Saks, and Ulrich Johann Airbus, Friedrichshafen, Germany Since several decades the scientific community works on concepts for a space based gravitational wave detector, such as the Laser Interferometer Space Antenna (LISA). With the spectacular first direct measurement of gravitational waves with an on-ground detector in 2015 [1] and the successful demonstration of the LISA technology package in the scope of the LISA Pathfinder Mission [2], the next phase of LISA technology development will start in The technology roadmap includes the development of suitable electric micro-newton thrusters which must be available in due time to avoid technical development risks for the mission. The thruster will be used for highly precise AOCS. Therefore, a detailed characterisation of the potential thruster candidates is necessary. A major difficulty is the verification in a low-frequency measurement band, specific to these applications. In particular, the direct measurement and the spectral characterisation of the thrust noise are crucial in determining the jitter performance. Technical Lead, Laboratory for Enabling Technologies, franz.hey@airbus.com PhD student, Laboratory for Enabling Technologies, max.vaupel@airbus.com 1

2 Nomenclature GAIA LISA AOCS FEEP RIT HEMPT COTS OLV ESC PSD = Global Astrometric Interferometer for Astrophysics = Laser Interferometer Space Antenna = Attitude and Orbit Control System = Field Electric Emission Propulsion = Radio Frequency Ion Thruster = High Efficiency Multistage Plasma Thruster = Commercial Off the Shelf = Oerlikon Leybold Vacuum = Electro-Static Comb = Power Spectral Density 2

3 I. Introduction Since several decades the scientific community works on concepts for a space based gravitational wave detector, such as the Laser Interferometer Space Antenna (LISA). With the spectacular first direct measurement of gravitational waves with an on-ground detector in 2015 [1] and the successful demonstration of the LISA technology package in the scope of the LISA Pathfinder Mission[2], the next phase of LISA technology development has started in The LISA Pathfinder mission demonstrated that as propulsion system for highly precise Attitude and Orbit Control cold gas thruster can be used. However, suitable electric micro-newton thrusters should be available in due time to in order avoid technical development risks for the mission with respect to the overall mass of the LISA spacecraft. Therefore, the intended technology roadmap includes the further development of electric micro-newton thrusters. For the development and qualification, a detailed characterisation of the potential thruster candidates is necessary. A major difficulty is the verification in a low-frequency measurement band, specific to these applications. In particular, the direct measurement and the spectral characterisation of the thrust noise are crucial in determining the jitter performance. Consequently, we developed our thrust balance specifically for this parameter regime and presented our first results at the IEPC 2013 and IEPC 2015 [3, 4, 7]. Since then, we have further improved the resolution, the stability and the reliability of the measurement facility as well as the measurement range of the balance. The actual thrust-balance generation is called Airbus micro-newton Thrust Balance MkIV. Interferometer Ref. Pendulum Meas. Pendulum Figure 1. Sketch of the double pendulum balance principle. Two fully symmetric hanging pendulums are combined with a laser interferometer as translation sensor. The combination of a highly precise and highly stable optical readout with the double pendulum setup leads to a common mode rejection of various noise sources and hence to a highly stable thrust measurement device. The MkIV thrust balance is still formed by a double pendulum assembly coupled with a highly precise and highly stable laserinterferometer, as it will be used on the LISA spacecraft. Figure 1 illustrates the balance principles. It consists of two hanging pendulums, a measurement pendulum and a reference pendulum. The laser beams between the interferometer and the pendulums are illustrated as thin (red) lines. The thrust measurement is performed by a differential translation measurement of the pendulums, i.e. the translation of the measurement pendulum minus the translation of the reference pendulum multiplied with the specific sensitivity results in the measured thrust. Therefore, external noise is not taken into account. This principle is called common mode rejection. The deflected pendulum is presented with dotted outlines. The common mode rejection applies not exclusively to seismic or thermo-mechanical noise, because the symmetry of the heterodyne laserinterferometer and the electronics used rejects different other noise sources such as laser frequency noise, electrical noise, etc. II. Airbus micro-newton Thrust Balance MkIV The MkIV thrust balance has consequently been developed based on the previous balance generations with the goal to further increase the stability, in particular despite a thermal load of an operational thruster. 3

4 Figure 2. The Figure presents the inner view of the vacuum tank to provide an overview of the complete measurement setup. At the top part the micro-newton thrust balance is visible, with the fully integrated thruster [5]. Additionally, the overall measurement range has been extended to up to 15 mn. The main goals of the development are: Higher stability Increased resolution Compatibility with higher device under test masses Wider measurement range More flexible thruster interface Material compatibility with alternative propellant, in particular Iodine An overview of the MkIV assembly is provided in 2, where the balance inside the vacuum facility is shown. The balance can be seen on the right side of the picture. The minor support structure is still made of commercial of the shelf standard profiles, whereas the main structure with the bearing is a custom designed highly stable Aluminium framework which is especially optimised in point of mechanical stiffness, but also for good handling and accessibility during thruster assembly integration. This is especially important because it is expected to perform thruster characterisation with various thruster candidates during the upcoming technology development phase of LISA. Another major design goal of the balance development was the compatibility with alternative propellants and especially to be compatible with Iodine. Therefore, all formally used materials has been analysed and replaced, if required, in order to be Iodine compatible. In parallel, the overall vacuum infrastructure has been adapted to be Iodine compatible. First results of the performed thruster firing tests with Iodine are presented in [6]. With respect to experimental tests which has been internally performed and a literature review all major system parts has been adapted. The turbo molecular pumps installed has been equipped with sealing gas connections. The fore stage pump has been replaced by a dry screw pump. A filter made of copper wool has been placed in the vacuum feed line between the screw pump and the turbo molecular pumps. It has been found that the cryo pump which is connected to the system, as described in [4, 5], has been already being able to be operated with Iodine. The facility still offers L/s of pumping speed. The thrust balance consists of two fully symmetric pendulums. A more detailed overview of a single pendulum is given in figure 3. The pendulum structure (in green) has been divided into three parts. A 4

5 Thruster Setup Thermal Shield Eddy Current Brake Voice Coil Counterweights Insulation Spacer lmirror Balance Support Structure Pendulum Structure Mirror Bearing Parts Electro Static Comb lvoicecoil lthruster Figure 3. Illustration of a single pendulum sub assembly which is rotated by 90 degree. On the pendulum structure (in green) all other required balance parts and the thruster that shall be tested are mounted. The figure illustrates the distances between the pendulum pivot and the thruster, the mirror and the voice coil. part where the pendulum is connected to the bearing (left side), a middle part where thruster equipment or calibration weights can be placed and a thruster mounting part where the thruster under test can be mounted. All parts are thermally insulated via two insulation spacers made of a thermally non-conducting ceramic. The middle and the thruster part can be adapted in order to be compatible with various thruster types, form factors or masses. The bearing structure (in grey) is formed by up to 20 leaf springs. The springs function as electrical connection of the thruster and other measurement equipment on the pendulum. The number of springs can be adapted onto the specific need of performed thruster tests e.g. if the investigated thruster requires a high input power, wider springs can be installed in order to reduce the thermal load of a single spring. The ceramic where the springs are mounted and which acts as reference plane has been replaced by Aluminium Nitride. This highly thermally conduction ceramic enables a homogeneous heat spread above the overall bearing and further reduces the thermal drift induced by heavy changes of the electrical loads installed. On the left part of the bearing, the conducting plates made of cooper and the required mirror are installed. The cooper plates are required for the eddy current brake which damps the first eigenfrequency of the setup. This is necessary to enable the common mode rejection as presented in [5]. The mirror is mounted onto piezoelectric steppers which allows an alignment of the laserinterferometer even in cases where the vacuum chamber is already evacuated. On the middle part of the structure a voice coil is placed on each pendulum, thus the balance is able to perform measurements in open- and closed loop. Typically the closed loop is used for direct thruster measurements and the open loop applies for highly precise thrust measurements. As mentioned, the thruster is placed on the bottom side of the pendulum. Above the thruster a radiation shield can be installed (bright blue). At the backside of the thruster assembly an electro static comb is mounted. With the electro static comb the calibration is performed and also the matching of the pendulums transfer functions is performed via the comb assembly. A matching of the transfer function is usually performed as part of the commissioning before measurements are performed. The matching increases the effect of the common mode rejection i.e. the stability and resolution of the assembly. Before a thrust measurement is performed the targeted sensitivity of the assembly have to be defined. Depending on the sensitivity the resolution, measurement range and bandwidth can be selected. The sensitivity of the pendulum can be described by S= E wtotal t3spring m2p g0 lcog lmirror ( + ) 2 lthruster 6 lspring lthruster Ipendulum, (1) where lmirror is the distance between the mirror axis and the rotation axis, lthruster is the distance between point of thrust and rotation axis, the E-Module of the springs is E, wtotal is the overall wide of all springs, tspring is the spring thickness, lspring is the acting length of the spring, mp is the mass of the overall assembly, g0 is the acceleration of the earth surface, the distance between the centre of gravity and the rotation axis is 5

6 called l cog and the inertial moment of the pendulum is I pendulum. Typical sensitivity values are in the order of 0.02 textmu N / nm. III. Measurement Results and Analysis After the full integration of the balance assembly different tests has been performed in order to characterise the balance and its performance Sensitivity, µn/nm Calibration Run Number Figure 4. Result of a long term calibration test of about 51.6 h where the sensitivity is plotted against the calibration runs. The variation of the pendulum sensitivity is less than 1.5 %. In figure 4, the result of a long term calibration test is presented, where the calibration run number is plotted versus the measured sensitivity. The squares are the measurement points with attached error bars. For the measurement 155 calibrations runs has been performed. Every 20 minutes a new run was started so that the presented data covers the variation of the thrust stand sensitivity in a time frame of about 51.6 hours. The variation of the pendulum sensitivity is less than 1.5 % and the mean value ( µn/nm) is inside the estimated uncertainty budget. The measurement underlines the stability of the whole thrust balance assembly. Moreover, the repeatability of the defined calibration process has been demonstrated. The determination of the specific transfer function of the balance setup and the impact of the eddy current brake was also part of the balance characterisation. An exemplary illustration of the frequency sweep is presented in 5, where the input current of the voice coil (solid line, in na) and the measured translation of the balance (dashed line, in nm) is plotted versus time (in s). Based on the measured data, a fourier transformation has been performed in order to assess the transfer function which are summarised in figure 6. The dotted curves and the dash-dotted curves are the Bode plot of a thrust balance setup where the eddy current brake was set to a lower influence by extending the distance of the permanent magnets to the conducing plates. Therefore, the eigenfrequency is not fully damped. It is visible at 0.7 Hz. The black solid curves present a measurement with fully activated eddy current brake. It can be observed that the natural frequency of the thrust stand is fully damped, the system is overdamped. Therefore, the response time of the pendulum is much lower, and the sensitivity at higher frequencies is reduced. The effect is also evident in the phase plot. The higher damping leads to a high phase shift which also would influence the PID control loop of the voice coil in closed loop operation. As mentioned before the thrust balance has been also adapted in order to be able to perform thrust measurements in the lower milli-newton range, a measurement example of this use-case is presented in figure 7. The thrust (in µn) is plotted versus the time (s), the red curve illustrates the calculated thrust and the blue curve represents the measured thrust. It can be seen that the calculated thrust and the measured thrust are fitting together. The visible noise of the calculated thrust is a result of the massflow controller used. The measurement underlines the ability of the test stand to perform precise real thrust measurements also within thousands of micro-newton beside the originally targeted application to measure sub-micro-newton thrust levels. 6

7 1,000 Input Current, na Translation, nm 1, ,000 Time, s Figure 5. Exemplary illustration of the frequency sweep used where the input current of the voice coil (solid line, in na) and the measured translation of the balance (dashed line, in nm) is plotted versus time (in s). The thrust balance has been also used to perform direct thrust measurements of a highly efficient multistage plasma thruster with Iodine as propellant. This measurements are presented in [6]. The balance performance in the sub-micro-newton measurement regime has been also characterised. Figure 8 presents the result together with other measurements. It shows a Power Spectral Density (PSD), where the thrust (in µn) is normalised to the frequency (in Hz) and logarithmically plotted. The LISA requirement is presented as black curve, the NGGM requirement is shown as blue curve, the measured thrust noise performance of the RIT µx is plotted in green, the in-house developed small highly efficient multistage plasma thruster is given in yellow and the measured balance performance in shown as red curve. The red curve illustrates that the developed balance is fully compliant with the targeted LISA requirement and offers a resolution in the nano-newton regime. IV. Conclusion A fourth generation of a highly precise and highly stable micro-newton thrust balances has been developed. The so called Airbus thrust balance MkIV strongly relies on the technologies of its precursors i.e. the balance is still formed by a double hanging pendulum setup coupled with a laserinterferometer as optical readout. The balance MkIV has been especially designed to further increase the stability of the balance, extend the measurement range to milli-newton thrust levels and to be compatible with alternative propellants in particular with Iodine. In parallel to this activities the overall test facility of Airbus in Friedrichshafen has been modified to be able to perform thruster tests with Iodine. The major design changes such as the insulation ceramic spacer in between the balance structure has been presented and an overview about the general mechanical assembly was shown. Measurements in order to characterise the new thrust stand has been performed and the results has been presented. The results illustrated that the overall long term stability of the setup is as excellent as predicted. A major part of the performed balance characterisation was the determination of the balance transfer function and the impact of the eddy current brake that is suppressing the first eigenfrequency of the setup. It has been demonstrated that the first eigenfrequency can be fully suppressed with the damper. Additionally, it has been experimentally verified that the transfer function can be adapted via simple changing the distance between the permanent magnets and the conduction plates. The balance has been also tested for thrust measurements in the order of thousands of micro-newtons, a representative measurement example has been provided. The test results of our first thrust measurement with Iodine as propellant are be presented in [6]. 7

8 Amplitude, µn 10 0 Trans. Fct. by ESC Partly damped Trans. Fct. by VC Partly damped Trans. Fct. by VC Fully damped Phase, deg Frequency, Hz 10 1 Figure 6. Example of the measured Bode plot of the thrust balance for different eddy current brake configuration. The measured response were applied with different actuators to demonstrate that the measurement is not dependent on the actuator used. In order to demonstrate the balance performance in the micro-newton regime also thrust noise measurements have been perform, with the result that the balance fulfils the LISA requirements and that it is feasible to perform thrust measurements in the nano-newton regime. References [1] B. P. Abbott et al., Observation of Gravitational Waves from a Binary Black Hole Merger, Physical Review Letters (2015). [2] M. Armano et al., Sub-Femto-g Free Fall for Space-Based Gravitational Wave Observatories: LISA Pathfinder Results, Physical Review Letters (2016). [3] Franz Georg Hey, et al., Development of a Highly Precise Micro-Newton Thrust Balance, IEPC (2013). [4] Franz Georg Hey, et al., Development of a Highly Sensitive Micro-Newton Thrust Balance: current status and latest results, IEPC (2015). [5] Franz Georg Hey, Development and Test of a Micro-Newton Thruster Test Facility and Micro-Newton HEMPTs, Springer (2017, to be published). [6] Max Vaupel and Franz Georg Hey, et. al. The Next Generation milli-newton µhempt as Potential Main Thruster for Small Satellites, IEPC [7] Franz Georg Hey, et. al. Development of a Highly Precise Micronewton Thrust Balance, Plasma Science, IEEE Transactions on 43.1 (Jan. 2015). 8

9 Figure 7. Thrust measurement at the operating point of 700 V anode potential, 2.5 sccm mass-flow and an average on-state anode current of 96.4 ma, the measured thrust was 2230 µn Thrust Noise, µn Hz -1/ Euclid & NGGM Requirement mini NG-µHEMPT LISA Requirement RITµX Full Balance Setup Frequency, Hz Figure 8. The plot presents the directly measurement thrust noise performances which has been presented in the thesis. The dashed curve illustrates the requirements of NGGM and Euclid. The dash-dot-dotted curve shows the requirement of LISA. The solid curve is the measured thrust noise performance of the developed mini NG-µHEMPT.The dash-dotted curve shows the measured thrust noise performance of the RITµX that has been tested with the developed micro-newton thruster test facility. The performance of the thrust balance which is part of the facility is presented as dotted curve. 9