A 2.4 GHz High Speed Communications System for Cubesat Applications

Transcription

1 A 2.4 GHz High Speed Communications System for Cubesat Applications Daniel G. Kuester, Pradeep Narayan P Radhakrishna Colorado Space Grant Consortium University of Colorado at Boulder Abstract This paper proposes a deployable solution and addresses design challenges for an S-band Microwave Cubesat communication system using consumer off-theshelf (COTS) components. Compared to typical existing amateur radio UHF systems on cubesats, this solution offers data throughput improvements, reduced mass, and reduced system-level and antenna deployment complexity by taking advantage of broader available bandwidth and shorter wavelengths at 2.4GHz. It also addresses challenges inherent to this band, such as antenna radiation irregularities, licensing, and pervasive ground interference. Successful tests of improved communications systems like this may help in obsolescing current UHF systems on future satellites, making a broader assortment of data-intensive missions possible. The new system will be demonstrated on the Hermes cubesat, which is expected to launch in Overview 1.1 Cal Poly s Cubesat The cubesat is an overarching satellite structure and deployment specification for a small 1 kg, 10cm x 10cm x 10cm picosatellite that was conceived by California Polytechnic State University (CalPoly). Cal Poly [1] provides specifications for standardized structural and design guidelines, a flight deployment system, and various forms of logistical support. Over sixty separate schools and several companies have made the cubesat a popular size, thanks in part to its cost-effective size and mass. This small, light form factor introduces key design challenges for most cubesats. The small surface area limits solar cell size and therefore total system power availability. The small size and mass also limits the number of devices which can fit into the satellite, and the size of deployable instruments and antennas. 1.2 Hermes Hermes [2] (or Cubesat I) is the first in a series of cubesats expected to be constructed and flown by the Colorado Space Grant Consortium (CoSGC). Its primary mission is as a prototype system bus by implementing first incarnations of typical core systems: power distribution fed by solar panels, command and data handling through a microcontroller, a passive attitude control system, and a 1200 baud UHF communication system. In addition, it has a secondary mission as the vehicle for our experimental S-band communication system. 1.3 S-band System Goal Amateur cubesats typically use cost-effective consumer off-the-shelf radios that, while proven reliable, rarely exceed data throughput greater than 1200 baud. This bottleneck makes missions involving generation of significant digital data such as photographs - unfeasible. As part of our attempt to alleviate this problem, Hermes will provide a bus for an experimental new communications system.

2 The communications system provides a 2.4GHz (12.5cm) link between the satellite in Low-Earth Orbit (LEO) and a suitably equipped ground station. Its primary goal is to demonstrate the technical feasibility of a high frequency communications link in orbit for future small satellite communications systems, and to upgrade the current ground station facility to support this kind of communications in future missions. A successful test will demonstrate data throughput greater than 9600 baud between the satellite and our ground station. 2.1 Modem 2. System Design The design (summarized in fig. 1) is split into two physically separate stations: the flight subsystem, which is built into the satellite, and the ground station, which will be constructed at a fixed site near CoSGC in Boulder, Colorado. These communicate via a linearly polarized (LP) link operating in the GHz (S-band) portion of the amateur radio service, between a 1.5m dish on the ground and a small ¼ wave (3.1cm) monopole in flight. MicroHard MHX2400 modems at both ends of the link transmit and receive digital data to and from processing hardware. Figure 1: System overview Figure 2: MHX 2400 Modem The Microhard MHX2400 [3] OEM modem module (figure 2) can operate between 2400 and 115k baud by modulating signals with Gaussian Frequency Shift Keying (GFSK). It is designed to connect to external logic circuitry through unshifted (0-5V TTL) RS-232, and to an antenna through a 50-ohm MCX connector. The board can transmit up to 1 Watt (30 dbm, or 0 db) of RF power, and has a receive sensitivity of -108 dbm (-138 db) with a bit error rate (BER) of 1 in This modem will be placed on both sides of the antenna link. On the ground, it will be connected to the ground antenna through its MCX RF connector, and an inexpensive consumer PC through the serial connection for data processing. On the flight side, the C&DH system controls the system through the serial link. The MCX connector makes the modem too long to fit properly in the structure, so it will be removed and replaced with an SMA connector, where the flight monopole will be attached.

3 Of the increasing array of available modems, the MHX2400 is ideal for our project because it provides well-documented (and data-transparent) configuration, a straightforward TTL control interface, improved data throughput, and efficient RF power conversion. Other modems use too much DC power, cost too much, or transmit too little RF power. In addition, we can use this modem on our ground station without any separate radio or transceiver equipment, because the modem transparently supports Doppler shift correction up to ±30 KHz, automatic channel selection, and frequency detection. 2.2 Flight Antenna The flight antenna presents a slew of difficulties. One unique to the S-band (compared with UHF systems) is that a 10cm satellite structure is electrically significant relative to the 12.5cm RF wavelength, and will distort radiation patterns. The cubesat itself imposes other; to accommodate attitude control limitations, the beam pattern must be as broad as possible to maximize ground communication duration for as many satellite orientations as possible. Further, a structural requirement forbidding protrusions from the cube surface implies that our antenna must be deployed after launch. (3.1cm) monopole. It will be mounted at the center of a face sharing the satellite s rotational axis, shown mounted on the satellite in figure 3. To verify that the antenna s interaction with the cubesat structure will not excessively deform the radiation pattern, we performed simulations, shown in figures 4 and 5. This simulation suggests that despite those interactions communication with the ground should be possible in the red and orange areas of the plot. Our structures team is currently working to design a deployment system to fulfill the remaining requirement for deployability. This is one circumstance which exposes a benefit of the smaller S-band wavelength: at 3.1cm long, the antenna is less than half the length of a cube face, and can easily be folded directly down onto the face of the cube before deployment. It will be lashed there to the structure using nylon wire before launch, and will released (using a magnetically-latched pin) by a signal from command logic 15 minutes after the satellite is ejected from the launch vehicle. Figure 3: Flight antenna, after deployment) Figure 4: 3-D monopole radiation pattern (strongest in red) To make the radiation pattern as broad as possible, the flight antenna will be a single quarter-wave

4 The mesh minimizes problems with wind loading on the mounting hardware caused by periodic local windstorms. To direct the antenna, we have selected a Yaesu G5500 azimuth-elevation (az-el) rotator. This will be controlled through a serial computer interface to steer the antenna beam toward the satellite, based on TLE data received by mission operators. 3. Operational Considerations 3.1 Link Power Calculations and Constraints Figure 5: Orthogonal radiation pattern (Nadir is up) 2.3 Ground station The 2.4 GHz extension to the ground station will consist of a new ground antenna, an antenna mounting and steering apparatus, the modem, and a low-noise amplifier (LNA) to correct for line losses. The various losses like atmospheric, free space, Ionospheric and rain losses are accounted for although rain and ionospheric losses are insignificant at 2.4 GHz. Atmospheric and free space losses vary in accordance with the distance between the satellite and the ground antenna which changes based on elevation angle. Pointing and polarization losses are taken into account using precise formulae. The LNA at the ground station is being chosen such that the transmission line loss is zero. The link margin varies between 10 db for the satellite vertically above (Elevation angle of 90 degrees) and 6 db for an elevation angle of 40 degrees with a baud rate of 57.6 kbps. Our simulations using Satellite Tool Kit (STK) has shown us that we should be able to achieve a data transfer of approximately 200 k Bytes per pass at an orbital height of 700 km. 3.2 FCC Licensing & Regulations Figure 6: Ground antenna radiation pattern The ground antenna is a highly directional HyperLink [4] 1.5m (30 dbi) linearly polarized (LP) parabolic mesh dish fed with a dipole at the focus. The MHX2400 was it is primarily designed to be able to operate in the ISM (FCC section 15) service between GHz, which limits it to 1W (30dBm) output power. Combined with our high gain ground antenna, we achieve an EIRP of approximately 30 dbm + 30 dbi = 60 dbm significantly higher than the (approximately) 36 dbm limit for ISM operation so we will instead

5 operate our system under the less restrictive amateur radio service (FCC section 97 [5]), which allows transmission of up to 1500W. The terms of this service stipulate that all parties using this system will need to be licensed amateur radio (HAM) operators, which will be an important concern for HAM operators. 3.4 Ground Interference Interference is a significant problem, as the amateur radio service shares the GHz band with the relatively unregulated ISM band. Broad deployment of devices which take advantage of the ISM service has resulted in widespread proliferation of internet and telephone noise through populated areas. Though our ground antenna has a very tight 5.8º main beam width, it will still receive these stray signals through its side and back lobes with isotropic gains of up to 5 db. Signals received in those lobes will be more powerful than the heavily attenuated signals from the satellite 700km away, and therefore be sufficient to interfere with effective modem operation. Tests with the modem shall determine its noise rejection efficiency. 3.4 Elevation Masking Boulder s location at the foot of the Rocky Mountains in Colorado presents a significant obstacle to the west. West-facing elevations from the CoSGC s home in the Discovery Learning Center (DLC) on the CU-Boulder campus in Boulder are blocked below approximately 15º by foothills rising 900m above the Boulder valley. 4. Future Work 4.1 Polarization Analysis In its current design incarnation, the ground antenna s linear polarization avoids the 3dB link loss incurred by communicating between circularly polarized (CP) and LP antennas. Unfortunately, the possibility for cross-polarization between transmit and receive antennas presents further complications. First, in an ideal case where the two antennas polarizations are perfectly aligned (co-polarized), then there will be no polarization loss between them. As the angle between the two antennas increases to 60 degrees, the loss increases to 3dB, the same loss as between CP and LP antennas. As the angle approaches cross-polarization at 90 degrees, polarization loss can exceed 20dB, which would prevent a successful communication link. Another problem which can arise from this LP LP configuration is that of Faraday rotation as the signal passes through the atmosphere. At UHF bands, Faraday rotation alone renders an LP LP system impractical; the signal can be unpredictably rotated up to and past cross-polarization. Near 2.4GHz, however, we expect typical polarization rotation of approximately 20 degrees, which is an angle well within the 60 degree tolerance. Once attitude control simulations have been completed, we will be able to determine whether the anticipated polarization losses will allow our LP LP configuration to function. The satellite will operate in a polar sun-synchronous orbit, resulting in approximately north-to-south passes over Boulder; we believe that this, added to the 20 degree Faraday rotation, will still be within a reasonable threshold below the critical 60 degree mark. If not, then we will modify the ground station antenna to be CP by adding extra elements as necessary. 4.2 Ground Station Site Selection Interference and elevation masking will be the primary criteria in choosing a final ground station site. Logistically, the ideal choice is near the current UHF station on the roof of CoSGC s home on the roof of the Discovery Learning Center on the CU- Boulder campus. If tests determine that local noise interferes with nominal operation of the MHX2400, however, a radio quiet site (such as one currently under consideration north of Boulder) will need to

6 be found. Further, potential sites west (and therefore nearer the mountains) will have higher elevation masks. Sites with masking above the 40º operating elevation will shorten communication pass durations. 4.3 After Hermes: Null Steering To circumvent interference concerns at the main CoSGC site, a next step could be the design of a null steering antenna. Null steering (or pseudo inverse beam forming) uses an antenna array to steer the main beam towards the desired user and nulls (zero amplitude lobes) towards any interference, increasing the effective signal to interference ratio (SIR). However, this requires direction of arrival estimation and beam forming. For beam forming, either the weighted sum or Dolph Chebyshev [7] algorithms could be used. To build this, additional antennas as well as variable phase shifters are needed. Thus, by adjusting the phase of individual antenna elements, the desired radiation pattern can be achieved. 5. Conclusions This system provides benefits which open up exciting new possibilities for future cubesat missions that take advantage of improved system and communications features: Greatly improved data throughput Transparent Doppler shift correction Built-in TNC, reducing complexity of in-house flight logic Only 2 flight components, reducing systems integration complexity Simplified antenna deployment Reduced antenna mass Efficient DC power consumption Hermes is currently expected to launch in Acknowledgements We would like to thank the following people for helping to make the project happen: Chris Koehler, for giving us all opportunities by keeping CoSGC running for so long; Brian Sanders, for encouragement and logistical support through the project s entire life cycle; James Gorman, for applying his simulation expertise to help our flight antenna Lisa Hewitt, Jim Love, and the entire Hermes Cubesat team, for all their hard work in making Hermes happen Dr. Al Gasiewski, for design advice The ground station team, for working with us to realize the ground station infrastructure updates 7. References [1] Cubesat Design Specification (CDS), Revision 9, Armen Toorian, Amy Hutputtanasin, media/documents/developers/cds%20r9.pdf. [2] Hermes Cubesat Mission Homepage, [3] MHX GHz Spread Spectrum OEM Transceiver, Operating Manual, Revision 1.56, April 7, 2006, Microhard Systems Inc. [4] 2.4 GHz 30 dbi High Performance Reflector Grid Wireless LAN Antenna - HG2430G, HyperLink Technologies, Inc. [5] FCC PART 97 Rules and Regulations Amateur Radio Service, [6] Adaptive Antennas for Wireless communication, Undergraduate Project ( ) by Pradeep Narayan P R, Prasad PBN, Prakash Natarajan. [7] Matrix Formulation for Dolph-Chebyshev Beam forming, A. Zielinski, Dec 1986, Vol 74, Issue 12.