A Data-Taking System for Planetary Radar Applications
JJanuary 12, 2021 2:1 pfs
Journal of Astronomical Instrumentation © World Scientific Publishing Company
A Data-Taking System for Planetary Radar Applications
Jean-Luc Margot , Department of Earth, Planetary, & Space Sciences, UCLA, Los Angeles, CA 90095, USA, [email protected] Department of Physics & Astronomy, UCLA, Los Angeles, CA 90095, USA, [email protected]
Received October 23, 2020; Revised November 17, 2020; Accepted November 24, 2020;Most planetary radar applications require recording of complex voltages at sampling rates of up to 20 MHz. Idescribe the design and implementation of a sampling system that has been installed at the Arecibo Observatory,Goldstone Solar System Radar, and Green Bank Telescope. After many years of operation, these data-takingsystems have enabled the acquisition of hundreds of data sets, many of which still await publication.
Keywords : Radar; Planets; Asteroids; Data acquisition.
1. Motivation
Planetary radar astronomy is a discipline that enabled major advances in our understanding of the scale ofthe Solar System, spin states and surface properties of planets and satellites, and the physical and dynamicalproperties of asteroids (e.g., Ostro et al. , 2000; Giorgini et al. , 2002; Margot et al. , 2002; Chesley et al. ,2003; Campbell et al. , 2006; Ostro et al. , 2006; Margot et al. , 2007; Taylor et al. , 2007). For reviews of thefield, see Ostro (1993); Campbell et al. (2002); Ostro et al. (2002); Benner et al. (2015). For an historicalaccount of the development of the field up to 1996, see Butrica (1996).The field is distinctive in part because essentially all results come from only two transmitting facilities,the Arecibo Planetary Radar and the Goldstone Solar System Radar. The facilities are complementary inthat the former is ∼
15 times more sensitive and the latter has access to a fraction of the sky that is ∼ et al. , 2016). The facilities also probe slightly different depths below the surface becauseArecibo transmits primarily at S band (2380 MHz) whereas Goldstone transmits primarily at X band(8560 MHz). Bistatic observations with reception at the 100 m Green Bank Telescope (GBT) or otherantennas also occur for a small fraction of the observations.Most planetary radar observations yield either 1D power spectra or 2D range-Doppler images (Evans &Hagfors, 1968), although 3D topographic maps can be obtained in some circumstances (Margot et al. , 1999).The sampling rate requirements for power spectra are relatively modest ( <
60 kHz) because the Dopplerbroadening of Solar System bodies at S band and X band do not exceed 30 kHz, with the exception ofSaturn’s rings (Nicholson et al. , 2005). For 2D imaging, the transmitted waveforms are most commonlyencoded with pseudo-noise (PN) binary phase codes or modulated with linear frequency (chirp) ramps(Margot, 2001). In both cases, the achievable range resolution is inversely proportional to the bandwidthof the transmitted waveform.NASA funded a major upgrade of the Arecibo Planetary Radar in the late 1990s, which included a1 MW S-band transmitter that can support 25 MHz bandwidth modulation (Goldsmith, 1996). In practice,7.5 m resolution images are obtained by encoding the transmitted waveform with a 0.05- µ s-baud PN code,corresponding to a bandwidth of 20 MHz. However, the data acquisition systems available after the upgradewere limited to sampling rates of 5 MHz, a factor of four lower than required for maximum resolution. This Corresponding author. 1 anuary 12, 2021 2:1 pfs Jean-Luc Margot situation prompted the development of a baseband recording system for use at Arecibo, GBT, and otherreceiving sites. This data-taking system was subsequently deployed at Goldstone as well.
2. Design Considerations
Radar echoes in two orthogonal polarizations are detected by low-noise receivers. The receiver signalsare amplified, filtered, and typically converted to baseband prior to digitization for computer storage andprocessing. Important design considerations include the resolution of the analog-to-digital (A/D) converterand the number of bits that are retained for subsequent processing. The data rate is given by d = 2 × n pol × n bits × f s , (1)where n pol is the number of recorded polarizations, n bits is the number of recorded bits, f s is the samplingfrequency in Hz, and the factor of 2 indicates that both the in-phase and quadrature components of thevoltage signal are sampled, i.e., a complex quantity. For 8-bit sampling of 2 polarizations at 20 MHz, thedata rate is 640 Mb/s or 80 MB/s, which was prohibitive after the Arecibo upgrade.Fortunately, the first processing step in planetary radar imaging is a range compression operationimplemented as a digital correlation. Hagen & Farley (1973) demonstrated that digital correlations can becarried out effectively with two-, three-, and four-level sampling of the input signals, resulting in surprisinglylow degradation of the signal-to-noise ratio (SNR) compared to that obtained with a finely quantized signal.For instance, the ratio of the power obtained with a two-bit (four-level) sampler to the power obtainedwith an ideal correlator exceeds 88% for a wide range of quantizer threshold settings (Kogan, 1998). Inparticular, setting the sampling thresholds at zero and ± { -3, -1, +1, +3 } (Schwab,1986). The design requirement was therefore established at 2-bit sampling of 2 polarizations at 20 MHz,or a data rate of 20 MB/s.The requirements for continuous sampling at low bit resolution are unlike those of most scientific orindustrial applications. Commercial data acquisition products typically prize sampling resolutions of atleast 8 bits and provide no straightforward mechanism for retaining only the most significant bits. Thisrealization led to a custom-built design and implementation for what became known as the Portable FastSampler (PFS), with available sampling modes listed in Table 1.mode channels bits sampling rate data rate0 2 1 (N/A)1 2 2 5 – 40 MHz 2.5 – 20 MB/s2 2 4 5 – 20 MHz 5 – 20 MB/s3 2 8 5 – 10 MHz 10 – 20 MB/s4 4 1 (N/A)5 4 2 5 – 20 MHz 5 – 20 MB/s6 4 4 5 – 10 MHz 10 – 20 MB/s7 4 8 (N/A) Table 1. Description of sampling modes.
3. Hardware Implementation
The sampling requirement is achieved with two Analog Devices AD9059/PCB, which are dual-channel 8-bitA/D boards with maximum conversion rates of 60 million samples per second. These boards require onlya +5 V power supply and a TTL-compatible encode clock. Full-scale on the A/D converters is achievedfor input signals of 1 V peak-to-peak driving a 50 Ω termination. Therefore, input voltages with standarddeviations near 0.25 V (1 dBm) yield optimum two-bit sampling. Digital outputs are TTL compatible.anuary 12, 2021 2:1 pfs
A Data-Taking System for Planetary Radar Applications In order to emulate 2-, 4- or 8-bit sampling, the 32 A/D digital outputs are connected to a pro-grammable logic device (PLD) for bit selection and packing. A device from the Altera MAX 7000S family(EPM7128SLC84-10) with plastic J-lead chip carrier (PLCC) 84-pin packaging is used for this purpose.It requires a +5 V power supply and accepts TTL-level input voltages. The supply voltages for internallogic and input buffers (VCCINT) and for output drivers (VCCIO) are both set to 5 V. The selected bitsare packed into sixteen data bits that are sent to four differential line drivers (SN75ALS194) for RS-422transmission over a 7-foot-long twisted pair cable (EDT CAB-AA 016-00427-00).A parallel interface card (EDT PCI CD-20) is used to transfer the data to computer memory andstorage disks at rates of up to 20 MB/s. The interface card is connected to the PCI bus of a rack-mountedcomputer that runs the Linux operating system. The interface card can handle four general-purpose controloutputs (FUNCT0–3), which are used to specify the sampling mode (written to FUNCT1–3) and to togglea data acquisition enable signal (written to FUNCT0).Two station clock signals phase-locked to a high-accuracy frequency standard are required. The firstsignal is a one pulse-per-second (PPS) TTL signal, which is used to trigger the start of the data acquisition.The second signal is a +13 dBm ( ∼
4. Software Implementation
A computer program ( pfs radar ) written in C can be executed from the command line or graphicaluser interface to control the operation of the data-taking system. Command-line options enable spec-ification of the sampling mode ( -m ), recording duration in seconds ( -secs ), and start time ( -startyyyy,mm,dd,hh,mm,ss ) specified in Universal Time. The computer clock is synchronized to accurate timeservers with the Network Time Protocol (NTP).Planetary radar observations require extreme ( <
10 ns) timing precision. Upon execution of the data-taking program, the computer allocates ring buffer memory, opens log and data files, then suspends exe-cution until 0.5 s before the desired start time. At that point, the data acquisition enable bit (FUNCT0)is activated, which signals to the PLD that data acquisition will commence on the upcoming rising edgeof the 1 PPS signal.When the rising edge of the 1 PPS is detected while the data acquisition bit is enabled, the PLDtoggles the Input Data Valid (IDV) bit, which instructs the 16-bit interface card to transfer data to PCIbus memory. The PLD is programmed to generate a Receive Timing (RXT) signal for use by the interfacecard, which stores inputs only at the rising edge of the RXT signal. Data are transferred 16 bits at a timeuntil the C program detects that the end time of data-acquisition has been reached. At that point, thedata acquisition bit is deactivated, ring buffers are cleared, the log file is flushed, and the whole processcan start over.Data files contain exclusively the raw data in packed format. Each data file name encodes the receiverstart time in the format datayyyymmddhhmmss . Ancillary information about sampling mode, buffer sizes,data rates, etc. are stored in the separate ASCII log file. Data-taking progress can be monitored remotelyby examining this log file.Most radar observations proceed with multiple transmit/receive cycles, where the timing of the receivecycles can be calculated in advance. Many observations use a simple script that repeatedly invokes thedata-taking program with the value of the next receive start time.Ancillary programs enable data inspection and analysis in near real time. One can display histogramsand statistics of the input data with pfs hist and pfs stats . Unpacking and downsampling of the data canbe done with the programs pfs unpack and pfs downsample . Spectral analysis is facilitated with pfs fft .The data-taking software and all ancillary programs are available on GitHub at https://github.com/UCLA-RADAR-Group/pfs.anuary 12, 2021 2:1 pfs Jean-Luc Margot
5. Applications
The data-acquisition system was installed at Arecibo (2 units), the Green Bank Telescope (2 units), andGoldstone (4 units). It has been used to acquire most of the radar echoes at Arecibo (2000–present),Goldstone (2001–2014), and Green Bank (2001–2017).The system was used for the first radar detection of a Solar System object at NASA’s Deep Spacetracking station (DSS-63) in Madrid, Spain. The asteroid 6489 Golevka was detected on 4 June 1999 withthe expected Doppler bandwidth and at the expected frequency. Observations of the same asteroid werealso obtained at Arecibo in 2003 and led to the first detection of the Yarkovsky orbital drift (Chesley et al. ,2003).Arecibo radar images of asteroid 1999 JM8 were obtained at 15 m resolution with the data-takingsystem on 1–9 August 1999. At the time, these images were the highest resolution images of an asteroidever obtained (Figure 1). This distinction was lost when the NEAR-Shoemaker spacecraft lowered its orbitsufficiently close to asteroid 433 Eros. Analysis of the radar data revealed that the asteroid has an effectivediameter of 7 km and a non-principal-axis rotation with a dominant periodicity near 7 days (Benner et al. ,2002).
Fig. 1. This sequence of 15 m resolution Arecibo radar images show asteroid 1999 JM8 on consecutive days.
Observations of the Moon with Arecibo transmitting at 2380 MHz and the 25 m VLBA antenna inSt-Croix receiving were conducted on 19–21 November 2000. The data-taking system was used to recordimages at 30 m resolution (Figure 2).The data-taking system enabled some of the first scientific observations at the Green Bank Telescope.Arecibo-GBT radar images of asteroid 2001 EC16 and Venus at 15 m and 150–300 m resolutions, respec-tively, were obtained on 24–26 March 2001. Because the round-trip light-time to the asteroid was 11 s,and because it takes several seconds to switch between transmit and receive modes, monostatic observa-tions would have limited the frequency resolution to the reciprocal of the ∼ et al. ,2000).Other notable radar observations from Arecibo include Saturn’s rings in 2001, 2001, and 2003 (Nichol-son et al. , 2005), binary asteroid 1999 KW4 at 7.5 m resolution in 2001 (Ostro et al. , 2006), and manyobservations of the Moon (e.g., Campbell et al. , 2007).The sampling system has been used in a Goldstone-GBT configuration to produce high-precisionanuary 12, 2021 2:1 pfs A Data-Taking System for Planetary Radar Applications measurements of planetary spin states. The observations revealed that Mercury has a liquid outer core(Margot et al. , 2007) and enabled a measurement of the size of its core (Margot et al. , 2012). They alsoenabled the first measurement of the spin precession rate of Venus and revealed that Venus exhibits length-of-day variations of tens of minutes (Margot et al. , 2021).
Acknowledgments
JLM is funded in part by NASA grants 80NSSC18K0850 and 80NSSC19K0870. I thank Jeff Hagen andJoseph Jao for software contributions. The data-taking system was initially funded by the National As-tronomy and Ionosphere Center.
References
Benner, L. A. M., Busch, M. W., Giorgini, J. D., Taylor, P. A. & Margot, J. L. [2015] “Radar Observations ofNear-Earth and Main-Belt Asteroids,”
Asteroids IV , eds. Michel, P., DeMeo, F. E. & Bottke, W. F. (Univ. ofArizona Press), p. 165.Benner, L. A. M., Ostro, S. J., Nolan, M. C., Margot, J. L., Giorgini, J. D., Hudson, R. S., Jurgens, R. F., Slade,M. A., Howell, E. S., Campbell, D. B. & Yeomans, D. K. [2002]
Meteoritics and Planetary Science , 779.Butrica, A. J. [1996] To See the Unseen , NASA SP: 4218 (NASA).Campbell, B. A., Campbell, D. B., Margot, J. L., Ghent, R. R., Nolan, M., Chandler, J., Carter, L. M. & Stacy, N.J. S. [2007]
IEEE Transactions on Geoscience and Remote Sensing , 4032, doi:10.1109/TGRS.2007.906582.Campbell, D. B., Campbell, B. A., Carter, L. M., Margot, J.-L. & Stacy, N. J. S. [2006] Nature , 835,doi:10.1038/nature05167.Campbell, D. B., Hudson, R. S. & Margot, J. L. [2002] “Advances in planetary radar astronomy,”
Review of RadioScience , ed. W. Ross Stone (URSI).Chesley, S. R., Ostro, S. J., Vokrouhlick´y, D., ˇCapek, D., Giorgini, J. D., Nolan, M. C., Margot, J. L., Hine, A. A.,Benner, L. A. M. & Chamberlin, A. B. [2003]
Science , 1739, doi:10.1126/science.1091452.Evans, J. V. & Hagfors, T. (eds.) [1968]
Radar Astronomy (McGraw-Hill, New York).Giorgini, J. D., Ostro, S. J., Benner, L. A. M., Chodas, P. W., Chesley, S. R., Hudson, R. S., Nolan, M. C., Klemola,A. R., Standish, E. M., Jurgens, R. F., Rose, R., Chamberlin, A. B., Yeomans, D. K. & Margot, J. L. [2002]
Science , 132, doi:10.1126/science.1068191.Goldsmith, P. F. [1996]
IEEE Potentials , 38.Hagen, J. B. & Farley, D. T. [1973] Radio Science , 775, doi:10.1029/RS008i008p00775. anuary 12, 2021 2:1 pfs Jean-Luc Margot
Kogan, L. [1998]
Radio Science ∼ astro/aotms/misc/chirp.pdf.Margot, J. L., Campbell, D. B., Giorgini, J. D., Jao, J. S., Snedeker, L. G., Ghigo, F. D. & Bonsall, A. [2021] submitted .Margot, J. L., Campbell, D. B., Jurgens, R. F. & Slade, M. A. [1999] J. Geophys. Res. , 11875,doi:10.1029/1998JE900047.Margot, J. L., Campbell, D. B., Jurgens, R. F. & Slade, M. A. [2000]
IEEE Trans. Geoscience and Remote Sensing , 1122, doi:10.1109/36.841991.Margot, J. L., Nolan, M. C., Benner, L. A. M., Ostro, S. J., Jurgens, R. F., Giorgini, J. D., Slade, M. A. & Campbell,D. B. [2002] Science , 1445, doi:10.1126/science.1072094.Margot, J. L., Peale, S. J., Jurgens, R. F., Slade, M. A. & Holin, I. V. [2007]
Science , 710,doi:10.1126/science.1140514.Margot, J. L., Peale, S. J., Solomon, S. C., Hauck, S. A., II, Ghigo, F. D., Jurgens, R. F., Yseboodt, M.,Giorgini, J. D., Padovan, S. & Campbell, D. B. [2012]
Journal of Geophysical Research (Planets) , E00L09,doi:10.1029/2012JE004161.Naidu, S. P., Benner, L. A. M., Margot, J. L., Busch, M. W. & Taylor, P. A. [2016]
Astron. J. , 99,doi:10.3847/0004-6256/152/4/99.Nicholson, P. D., French, R. G., Campbell, D. B., Margot, J.-L., Nolan, M. C., Black, G. J. & Salo, H. J. [2005]
Icarus , 32, doi:10.1016/j.icarus.2005.03.023.Ostro, S. J. [1993]
Rev. Modern Phys. , 1235, doi:10.1103/RevModPhys.65.1235.Ostro, S. J., Hudson, R. S., Benner, L. A. M., Giorgini, J. D., Magri, C., Margot, J. L. & Nolan, M. C. [2002]“Asteroid radar astronomy,” Asteroids III , eds. Bottke, W. F., Cellino, A., Paolicchi, P. & Binzel, R. P. (Univ.of Arizona Press), p. 151.Ostro, S. J., Hudson, R. S., Nolan, M. C., Margot, J. L., Scheeres, D. J., Campbell, D. B., Magri, C., Giorgini, J. D.& Yeomans, D. K. [2000]
Science , 836, doi:10.1126/science.288.5467.836.Ostro, S. J., Margot, J. L., Benner, L. A. M., Giorgini, J. D., Scheeres, D. J., Fahnestock, E. G., Broschart, S. B.,Bellerose, J., Nolan, M. C., Magri, C., Pravec, P., Scheirich, P., Rose, R., Jurgens, R. F., De Jong, E. M. &Suzuki, S. [2006]
Science , 1276, doi:10.1126/science.1133622.Schwab, F. [1986] “Two-bit correlators: Miscellaneous results,” VLBA Correlator Memo 75, National Radio Astron-omy Observatory.Taylor, P. A., Margot, J. L., Vokrouhlick´y, D., Scheeres, D. J., Pravec, P., Lowry, S. C., Fitzsimmons, A., Nolan, M. C.,Ostro, S. J., Benner, L. A. M., Giorgini, J. D. & Magri, C. [2007]
Science316