Skip to Main Content
This paper describes the operating principles and the design of the Mars shallow radar sounder (SHARAD), an HF sounding radar devoted to the mapping of sub-surface features of Mars and currently operational on board the NASA/JPL's Mars reconnaissance orbiter spacecraft. Compared to its predecessor MARSIS, currently operating from ESA's Mars Express, SHARAD is characterised by an higher carrier frequency (20 MHz vs a max of 5 MHz for MARSIS) and a much wider signal bandwidth (10 MHz vs 1 MHz of MARSIS). This allows SHARAD to achieve a finer range resolution (15 metres unweighted in free space) at the expenses of ground penetration, which makes the instrument ideal to probe the shallow subsurface layers (to depths of hundreds of meters) which cannot be resolved from the surface by the much far- reaching MARSIS. SHARAD uses a 85 usec chirp signal with a PRF of 700 Hz and a peak power of 10 W, and radiates by means of a 10 meters fiber foldable tube (FFT) dipole antenna, with a wide-band matching network in charge of impedance matching with the transceiver. The most challenging requirement(especially considering the large fractional bandwidth of the system) is the level of the range sidelobes, which shall be below -55 dBc after the 6th lobe, to allow proper detection of the weak subsurface echoes in presence of strong surface returns. On this side, the design takes advantage from the wide download bandwidth made available by the MRO Spacecraft to keep the on-board processing to a minimum level (basically, only a programmable coherent presuming), and leave most of the processing (range compression and synthetic aperture) on ground. In this way it is easy to use Tx chirps and Rx transfer functions characterised on- ground as reference for range correlation, with the range sidelobes limited, basically, only by the stability of the RF hardware. The limited amount of on-board processing also helped in limiting the complexity of the instrument design and, therefore, its ma- s and power consumption. SHARAD uses a very simple architecture, with the transmit chirp generated directly on the RF frequency (using a digital chirp generator) before being amplified to the transmit level by a class C amplifier. The receiver is even more essential, providing direct amplification of the received signal (with programmable gain and band filtering) to an A-to-D converter operated in downsampling mode by digitising the signal at 26.6 MHz rate. In this way, the complete 10 MHz signal bandwidth can be represented unambiguously with only an acceptable amount of oversampling (30%) minimising the required hardware. Instrument control and processing tasks are performed by the same AD-21020 DSP (with the help of a couple of FPGAs). The presuming can be varied from 1 to 32 in powers of two steps, and the resolution of science data can be selected to be 4, 6 or 8 bits, to allow optimisation of the data rate vs the operating scenarios. The receive window position can either be controlled in open loop, using an a priori knowledge of S/C orbit and surface topography (which demonstrated to be a very robust approach) or in closed loop.