Nano-Satellites for Micro-Technology Pre-Qualification: The Delfi Program of Delft University of Technology R.J. Hamann, C.J.M. Verhoeven, A.A. Vaartjes, and A.R. Bonnema Abstract The Delfi program run at Delft University of Technology has the objective to provide a means of fast and affordable pre-qualification of new (micro-)technology in space. A second objective is to offer to MSc students the possibility to graduate on real hardware and software development for space missions. The first satellite in this program is Delfi-C3, a 3 kg CubeSat, which flies a new type of Thin Film Solar Cells, two Autonomous Wireless Sun Sensors and a miniaturized UHF-VHF transponder. Next in the series will be a Delfi-C3 successor, on which a more efficient power subsystem and a high efficiency Advanced Transceiver will be flown. Further missions are considered which will fly micro-technology now under development in the MISAT research program. The paper describes the Delfi- C3 mission and satellite with its payload. Emphasis is given to technical solutions and the way the project is implemented in the university s educational program. An overview is given of the current status of the project and plans for the future development of Delfi missions are outlined. 1 Introduction Delfi-C 3 [1] is the first in a series of nano-satellites of the Delft University of Technology. The satellite has been developed by students and staff of the faculties of R.J. Hamann Delft University of Technology, Faculty of Aerospace Engineering, Space Systems Engineering, P.O.Box 5058, 2600 GB Delft, The Netherlands e-mail: r.j.hamann@tudelft.nl, info@delfic3.nl C.J.M. Verhoeven Delft University of Technology, Faculty of Aerospace Engineering, Space Systems Engineering, P.O.Box 5058, 2600 GB Delft, The Netherlands A.A. Vaartjes Delft University of Technology, Faculty of Aerospace Engineering, Space Systems Engineering, P.O.Box 5058, 2600 GB Delft, The Netherlands A.R. Bonnema Delft University of Technology, Faculty of Aerospace Engineering, Space Systems Engineering, P.O.Box 5058, 2600 GB Delft, The Netherlands R. Sandau et al. (eds.), Small Satellites for Earth Observation, 319 C Springer Science+Business Media B.V. 2008
320 R.J. Hamann et al. Aerospace Engineering and of Electrical Engineering, Mathematics and Computer Science. It will act as a technology test-bed for three payloads: Thin Film Solar Cells (TFSC), an Autonomous Wireless Sun Sensor (AWSS) and a radio-amateur, nano-satellite-based UHF/VHF linear transponder. The project was initiated in November 2004. Based on a three-unit CubeSat, Delfi-C 3 will be launched from a standard CubeSat deployer such as the X-POD. Delfi-C 3 will make use of the amateur radio frequency bands for its telemetry downlink and tele-command uplink. Software will be made available to participating amateur radio operators, which will allow them to decode, display and send the telemetry to the central ground station in Delft. The satellite will not include an active attitude control system. A passive rotation rate damping will be used to limit its rotation rate. Furthermore, as the primary payloads do not require operations in eclipse a battery has been avoided, which simplifies the spacecraft design. A piggyback launch in September 2007 has been arranged on a Polar Satellite Launch Vehicle (PSLV) from India. It will be injected in a circular 98 inclination Sun-synchronous orbit at 630 km altitude. After launch and deployment, three months of gathering payload data is foreseen, after which the satellite will be switched into linear transponder mode. Figure 1 shows an exploded view of the Delfi-C 3 satellite. Fig. 1 Delfi-C 3 satellite (exploded view) At the date of writing Delfi-C 3 subsystems have been produced and are under test. System integration has started and functional and environmental tests are scheduled for completion before July 2007. 2 Technology Payloads The TFSC (Fig. 2) payload is based on the latest development in the area of photovoltaic cells by the company Dutch Space. The cells consist of a Copper Indium Gallium diselenide (CIGS) photovoltaic layer, which is deposited by evaporation
Nano-Satellites for Micro-Technology Pre-Qualification 321 on a titanium base layer of 25 micrometers. The cells will be integrated tile-wise, ensuring a minimal loss of active cell area. The aim of this new type of solar cell is to create a lightweight and low-cost product for future space applications. The target is a 50% cost reduction of solar arrays, while improving the power to mass performance with 50% compared to conventional solar cells. The efficiency will be more than 12%. Measuring the characteristic current-voltage curves and cell temperature under varying Sun input will test the performance of the TFSC payload. As the temperature of the thin foil cannot be measured directly, it will be measured by determining the electrical resistance of a dummy titanium cell mounted close to the actual TFSC s. The TFSC s cannot be body mounted because of their fragility, but also because of the high cell temperatures that would result from such a configuration. For this reason the four pairs of TFSC s are mounted at the tip of each deployable solar panel. Fig. 2 TFSC payload The AWSS (Fig. 3) is a development of TNO in the Netherlands [2] and is placed at the top and bottom of Delfi-C 3. The sensors will have a half-sized GaAs solar cell as their own power supply, making them independent of the satellite s electrical power system. This autonomy is accompanied by a wireless radio frequency link, using an adapted commercial off-the-shelf transceiver. The UHF/VHF transponder, which doubles as a transceiver in science mode, is of conventional design. It is however, for the first time that such a transponder, allowing five simultaneous users, will fly on a CubeSat. Telemetry, both science and housekeeping data, is broadcasted continuously and may be received with a standard radio-amateur ground station. The satellite is
322 R.J. Hamann et al. Fig. 3 AWSS payload operated from the main ground station at the Delft University of Technology, and has a total of 25 minutes ground contact time in three passes per 24 hours. To increase the amount of data received, a distributed ground network has been created composed of radio-amateurs worldwide. By means of software made available to them by the Delfi-C 3 project they will be able to receive and decode the telemetry and to send the raw and decoded data to the operations center in Delft, where they are processed and made available to the users Dutch Space, TNO and, of course, the Delfi-C 3 project itself. 3 Technical Constraints Most important constraints for a nano-satellite are size, power and mass. The size limitation drives towards the application of surface mounted devices (SMD s) wherever possible and requires the implementation of deployment mechanisms for e.g. the eight antennae, which have in this case a maximum length of 0.5 m, which should be accommodated in an envelope with a maximum length of 0.3 m. So a dedicated design using a tape spring (measurement tape) that allows to store each antenna in a volume of 10 40 50 mm has been developed (see Fig. 4). The Delfi-C 3 architecture originally has been designed as a star architecture with one, single central computer. However, this led to a system bus of more than 120 wires running from top to bottom through the satellite, consuming 18% of the total mass budget. The transition to a decentralized Command and Data Handling Subsystem (CDHS), where each print is controlled by a dedicated microcontroller resulted in a mass saving of 11% (star architecture: 0.530 kg, 18% of total; decentralized architecture: 0.214 kg, 7% of total). The transmitter uses most of the power (2 out of 3 Watts available). It then became obvious that conventional over-current protection (OCP) of the individual prints and subsystems would consume too large a part of the remaining power to allow satisfactory satellite operations. The solution was the introduction of a decentralized Electrical Power Subsystem (EPS), where a separate controller is used to protect and switch each print or subsystem. That way a power saving (exclud-
Nano-Satellites for Micro-Technology Pre-Qualification 323 Fig. 4 Modular antenna box (MAB) ing transmitter power) of 11% was achieved (conventional switch and OCP design: 0.182 Watt, 18% of total; decentralized switching micro-controllers: 0.073 Watt, 7% of total). The resulting architecture using no less than 18 micro-controllers is shown in Fig. 5. 4 Project Implementation in the Educational Program The project team currently comprises about 25 students, who are in their last year of studies; half of them are in Aerospace Engineering, and half in Electrical and Computer Engineering. Although the project originally has been started with MSc students only, there is currently about equal participation of MSc and BEng students. It is felt that this is an ideal team composition in the later phases of the project. The work the students do on Delfi-C 3 is part of their final thesis work. Six staff members are involved in the project for in average 50% of their time. A self-managed student team has run the first year the project. As the students were very committed to building this first Dutch university satellite, this has worked very well. Also the fact that two well-known space companies are customers for the project was very stimulating and offered great learning opportunities for the students involved. However, the transition from a paper study to more serious hardware bread-boarding has been made too late and appeared to require more guidance by staff, in particular by staff with industrial experience. Student tasks need then to be specified more accurately, and achieving timely technology readiness and maintaining the schedule must get priority over the learning by making mistakes, which is the normal operating mode in a student run project. This certainly is the case, when a commitment for a launch date has been made. In this phase students have to execute the tasks that are most urgent, even if they are normally considered to be too menial for MSc or BEng level graduation work or
324 R.J. Hamann et al. Solar Panels Solar Panel #1 SP2 SP3 SP4 InterConnect Board Z+ Measurement Board Z+ Electrical Power Subsystem Antenna Deployment µcontroller1 ADP2 MeBo EPS µcontroller1 MeBo Dataproc. µcontroller1 EPS Management µcontroller Auton.Wireless Sun Sensor Z+ Combination Board AWSS AWSS µcontroller ComBo EPS µcontroller RAdio Platform 1 Auton.Wireless Sun Sensor Z- AWSS FM430 12 V regulated power RAP EPS µcontroller1 RAP Bitshaper µcontroller1 RAP µcontroller1 On-Board Computer RAP2 REP2 RBP2 ICB Z- MeBo Z- RCP2 ADP3 ADP4 MEP2 MDP2 Fig. 5 Delfi-C 3 controller architecture (light blue boxes: decentralized CDHS; green boxes: decentralized EPS)
Nano-Satellites for Micro-Technology Pre-Qualification 325 if they are not directly in their field of expertise (it is, however, our firm opinion that this offers many unique learning moments to the students). Also, the demands on the university s infrastructure increase in this phase: Workshop capacity, delivery times, fast design-to-test iteration cycles, specific production expertise are key requirements which are often difficult and sometimes even impossible to meet in an academic context. Good relations with established electronic and space industries are essential to achieve the technical project goals, and, for certain tasks, a direct use of their facilities and expertise should be made. Finally, the use of Delfi-C 3 for educational purposes is not limited to the higher education institutions directly involved in the project. Thanks to the mission architecture selected it is very well possible to equip secondary schools with a relatively simple and cheap ground station (kit), that allows the students to e.g. receive the satellite data, to make these visible and execute ground-based reference experiments, to carry out satellite ranging experiments and to explore the fundamentals of orbital mechanics. Activities in this sense have been started and help to rise the interest in engineering sciences with the younger generation. 5 Current Project Status Delfi-C 3 is currently in the test and verification phase. Subsystem testing has been completed for the most part and integration testing has also started. Subsystem testing began in January 2007 with the Combination Board (ComBo). The ComBo is the interface of the FM430 On-Board Computer (OBC) board with the rest of the satellite and it houses the AWSS receiver. Problems discovered required repairs, which rendered the first ComBo as non-flight. A second ComBo has been tested and will be flight hardware. The two InterConnect Boards (ICB s) house the antennae and its microcontrollers manage the deployable mechanisms: two solar panels and four antennae. Tact switches monitor the deployment status and a simulation board was made to simulate the mechanisms during functional testing (Fig. 6). For the moment, simulated deploy times have been used to test the functionality. Actual deployment times will be implemented following the deployment tests. The two Measurement Boards (MeBo s) perform the measurements of TFSC payload. Initial problems with the circuits have been resolved and the board is now being functionally tested. Tests involving the Voltage Controlled Oscillator, which will be used as a back-up in case of OBC failure, are scheduled to be conducted anytime now. Version 2.0 of the OBC software is almost completed and OBC subsystem interface testing is underway. Design and development of the Radio Amateur Platform (RAP) proved to be quite an effort, but an elegant breadboard of the RAP has been completed and the flight PCB is expected to come in soon. Ground support equipment for the assembly, integration and test phase is being build. A stack dummy, for fit-checking subsystem boards, and the integration jig
326 R.J. Hamann et al. Fig. 6 ICB test setup Fig. 7 Delfi-C3 chassis with top and bottom panel in integration and test jig have just been completed. Figure 7 shows the integration jig that will also support the satellite during thermal-vacuum testing. At this moment work is focused on the ground system. The first version of the telemetry decoding software has been produced and tested. Work on the science and housekeeping data post-processing system is progressing well. The User and Software Requirement Documents are being compiled and architectural and detailed design and coding will be started next month. The Delfi ground station has been used already for several other missions (e.g. SSETI Express) and is being upgraded to full redundancy (one of the lessons learned of that mission).
Nano-Satellites for Micro-Technology Pre-Qualification 327 Much has been learned from the test and verification phase so far. Originally the idea was to have a proto-flight approach, but the problems encountered during the first subsystem tests of the Combination Board in January quickly proved that was not feasible, even for a small project like Delfi-C 3. The test and verification phase was for Delfi-C 3 the first moment during the project in which the design was properly confronted with real life practice and it became clear that (elegant) breadboards should have been built earlier in the project, if short development times of about two years are targeted for. Getting concept and ideas out in practice could have reduced the number of problems that were encountered during the beginning of the test and verification phase considerably. This is something that definitely will have to be done in the projects that follow. Another example of the failure of the proto-flight approach is the fact that the Modular Antenna Boxes had to go through five design and test cycles before a satisfactory performance was achieved. For more details on the verification program of Delfi-C 3 see [3]. 6Delfi-C 3 FOLLOW-ON Mission The spare of the Delfi-C 3 satellite possibly will be used for a follow-on mission. Modifications for this mission will be made in the Electrical Power Subsystem and the Communication Subsystem. The DelfiC 3 satellite has, like many other satellites, different operation modes that require different ways of signal coding and transmission. A specific transceiver optimal for one of the operation modes may be suboptimal for another mode. As CubeSats are too small to carry separate optimal transceivers for each mode, an elegant solution to this problem is the reconfigurable transceiver. In such a transceiver the OBC takes many decisions during the mission that are normally taken in design phase by the designer and remain fixed during the rest of the operational life. This means that, for example, the architecture of the transceiver is not fixed, but designed and implemented by the OBC based on the specific transmission needs, like data rate, modulation scheme and the environmental circumstances like available power and quality of the transmission path. Since it is possible to implement most of the transceiver as a system-in-a-package (SIP), which is a micro-electronic system consisting of several integrated circuits (IC s) bonded together and put into a single IC-package, the volume and mass penalty for having such an Advanced Transceiver (ATRX) onboard is negligible. The power consumption is never more than exactly needed to offer the required quality of service, since the OBC can always implement the optimal transceiver architecture. Delfi-C 3 has transmission modes in which the amplitude of the transmitted signal carries information and modes in which the information is contained in phase or frequency of the signal only. A first step into the direction of a fully reconfigurable transceiver is the design and implementation of a reconfigurable power amplifier (PA). Due to the limited onboard resources, the efficiency of the PA is of extreme importance. The antennae should transmit the maximum power available in the satellite.
328 R.J. Hamann et al. The core of the PA in the ATRX is a switching amplifier. The power from the solar panels is put directly to the antennas via high efficiency switches that can modulate the phase and the frequency of the transmitted power. This is a very high efficiency mode with efficiencies (theoretically) much better than 90%, but unfortunately without the possibility for amplitude modulation. In this mode the PA can only be used to transmit frequency/phase modulated telemetry data. In transponder mode, also amplitude modulation must be possible. To achieve this, a high efficiency DC/DC converter modulates the power from the solar panels with the information. This amplitude modulated power supply is then fed to the switching PA that puts it to the antennas at the right carrier frequency and that can add frequency or phase modulation if required. When a very high linearity is required, negative feedback will be applied to the PA-DC/DC converter system. Negative feedback is the most powerful means to linearize a system. Normally the stability of the feedback loop around a complex amplifier system might be questionable, but in the (also electrically) small and welldefined SIP-module, stability can be guaranteed. Figure 8 shows a diagram of the PA system of the ATRX. The mode-selection switches are used by the OBC to put the PA in the optimal mode for a specific transmission, varying from highly linear with still a very good efficiency to maximum power output without amplitude modulation at an extremely high efficiency. S2 AM Input S3 DC/DC S1 Detector Supply Detector Source Fig. 8 Power amplifier system of ATRX Class E/F PA Load In this follow-on mission also a more advanced (maximum power point tracking) EPS will be flown, where the excess power will be used to maximize the transceiver performance. Actual realization of this mission depends on the success of the Delfi- C 3 launch and operations and on the success of identifying candidate technologies from the MISAT program (see section 7) for a 2009 mission. 7 MISAT Derived Missions Within the MISAT program [4], see Fig. 9, a number of micro-technologies are developed which are potential candidates for next Delfi missions. The AWSS is
Nano-Satellites for Micro-Technology Pre-Qualification 329 already a pre-development of the Micro Sun Sensor, and the RF Front End and Smart Power technology will be incorporated in the Delfi-C 3 follow-on mission. For a next mission micro-propulsion and Micro Navigation System components may for example be combined in a simple formation-flying mission. Another possibility is to develop a nano-satellite attitude control system based on a micro-inertial platform. In the second half of 2007 a project will be started to identify candidate technologies for this third Delfi nano-satellite. Actual choice will depend the availability of flight-ready prototype technology end of 2009. MISAT Cluster Spacecraft Bus Spacecraft Payload Satellite Architecture Distributed Systems Micro Navigation System Triple frequency GPS receiver Micro Satellite System Arch. Auton. Formation Flying & Control Micro Sun Sensor Relative Navigation Reduction of size, mass, power Multi-sensor RF Front End Micro Cooler Software System Architecture Micro Propulsion Bus sensors and interfaces Micro Accelerometer Damage-tolerant Aluminium alloys Control of distributed systems Precision docking Smart Power Thruster Test Facility Fig. 9 Micro-technology development in the MISAT research program References 1. W.J. Ubbels, F.A. Mubarak, C.J.M. Verhoeven, R.J. Hamann, G.L.E. Monna, The Delfi-C 3 Student Nanosatellite an Educational Test-Bed for New Space Technology, Proceedings of the AMSAT UK 21th Annual Colloquium, Guildford, UK (2006). 2. J. Leijtens, K. de Boom, Small sensors for small (and other) satellites, Proceedings of the 6th IAA Symposium on Small Satellites for Earth Observation, Berlin, Germany (2007).
330 R.J. Hamann et al. 3. A.A. Vaartjes, R.J. Hamann, R. Amini, Integration and Verification of a Command and Data Handling Subsystem for nano-satellite projects with critical time constraints: Delfi-C 3, Proceedings of the 58th International Astronautical Congress, Hyderabad, (2007). 4. W. Jongkind, The Dutch MST Program MicroNed and its Cluster MISAT, Proceedings of IC- MENS 2005, Banff, Canada (2005).