Lunar and Planetary Science XXXVI (2005) 2292.pdf

SUMMARY OF ROCK ABRASION TOOL (RAT) RESULTS PERTINENT TO THE EXPLORATION ROVER SCIENCE DATA SET. P. W. Bartlett1,2, L. E. Carlson1, P. C. Chu1, K. R. Davis1, S. P. Gorevan1, A. G. Kusack1, T. M. Myrick1, and J. J. Wilson1, and the Athena Science Team, 1Honeybee Robotics; 460 W. 34th St., New York, NY 10001; [email protected].

Introduction: The Rock Abrasion Tool (RAT) The abrasion area surface finish varies significantly serves as the sample preparation device on the Mars with rock type. Generally, conventional grinding pro- Exploration Rover (MER) science payload. As an duces surface roughness heights less than 1.5 microns. end-effector on the rover’s robotic arm the RAT [1] Depending on rock types, grinding artifacts are at grinds a circular spot on the order of millimeters deep times evident, in the form of arcing scratches or swirls into a rock face to remove surface fines and weathered corresponding to the grinding wheel path of motion. layers, preparing the rock for imaging and spectral observations. As of 5 January 2005, 12 grinding op- Natural face erations have been performed at Crater by and 22 at by . Basic Abrasion Properties: Once the robotic arm Abrasion wall places and loads the RAT onto a rock surface, the Abraded face RAT affects high rotational speed face grinding. A grinding bit with two diamond grain embedded pheno- Inner pile- lic resin pads rotates Natural fracture near 3000 RPM and up ring feeds slowly in a Outer pile- circular path across up ring Unabraded the rock, creating the depression larger abrasion area. The 45.5 mm outer Fig. 3: Microscopic image from Spirit of abrasion of diameter of the abra- ‘’ Courtesy NASA/JPL sion is a fixed aspect Fig. 1: Diagram of grinding Cuttings. During and after grinding the RAT of the mechanism. wheel path clears cuttings from the abrasion area with brushes. A The overall depth of penetration varies with opera- brush bit moves with the grinding bit clearing the im- tion due to rock properties and available rover energy, mediate abrasion area and a long perimeter brush typically no greater than 10 millimeters. The depth of pushes the accumulated ring of cuttings away from the penetration also varies across the abrasion area due to abrasion area if it forms. Cuttings are prominent in the rock’s original surface geometry and the RAT’s post-grind imaging primarily as an even coating on the approach angle. Abrasion areas can be less than fully rock face nearby and as accumulated rings surrounding circular and can include irregular unabraded areas in- the abrasion area. ternally. Often an overall depth is reported with a Particle size analysis [2] performed on cuttings grind operation, describing the distance from the high- produced by grinding at 7 torr Earth atmosphere into a est point on the original rock face within the circular moderately hard Palisades basalt shows that roughly grind area to the floor of the abrasion. 91% of the cuttings by weight are of 5 – 90 microns in size. The size fraction less than 5 microns is repre- Scrape mark caused by RAT sented as 7 wt% and the fraction greater than 90 mi- contact mechanism (rare) crons is represented at 2 wt%. Rock types, grind depths, and grinding orientations Pile-up rings Spots masked from with respect to gravity can effect the clearing of cut- of cuttings cuttings by RAT tings, leaving clumps or a coating on the abrasion contact mechanism floor. The impact of subsequent in situ sensor obser- vations can be apparent in the cuttings. Marks left by Thin coating of the Mössbauer spectrometer and APXS contact mecha- cuttings Abrasion area nism plates are frequently evident as they compress Fig. 2: False color Pancam composite from Spirit of cuttings or move cuttings into the hole. Also, a abrasion of ‘Adirondack’ Courtesy NASA/JPL/Cornell roughly 2 millimeter hole is sometimes made in col- Lunar and Planetary Science XXXVI (2005) 2292.pdf

lected cuttings by the Microscopic Imager’s surface properties, with respect to a grinding operation. De- contact sensing rod. convolving indications of separate standard rock Actuator Telemetry: The RAT employs three ac- physical property values such as that of shear strength tuators to affect grinding in a planetary arrangement. from this bulk property is non-trivial, however possi- Rotation of the bits about their centers, orbiting of the ble, in conjunction with supporting rock observations bits about the their common center, and plunging of and ground testing for comparison. The standard the bits into the rock are each independently actuated. physical property most closely linked with grindability The orbit motion is under closed loop control reacting is hardness. to current draw by the bit rotation motor, effectively Homogenous rock composition makes physical varying the orbit aggressiveness based on torque at the properties deductions much more straight forward. bit. The plunge motion is performed in discrete steps Rocks comprised of multiple materials of different once or twice per orbit revolution. Data from the three physical properties, especially with variation on spatial actuators comprise the bulk of the RAT telemetry. scales relevant to grinding, complicate analysis, requir- Time histories of motor position and current are re- ing imaging and spectral analysis even more so for corded at a rate of either 1 or 2 Hz. Temperature as interpretation. measured by a sensor mounted on the plunge actuator Spherules at Meridiani: The rocks encountered and switch states of the contact mechanism are also by Opportunity containing spherules or “blueberries” included in the telemetry, as well as software states. created special cases in RAT performance. In most Rock Physical Properties: Since the RAT per- cases the spherules require significantly more grind forms a mechanical operation on a rock, deductions energy than the base rock and are evident in the RAT can be made about the rock’s physical properties. The telemetry. Once the base rock is ground away and a energy required to grind away the rock is calculated, spherule is ground down to nearly hemispherical in normalized by volume of rock removed, and inter- shape, it is commonly unseated and thrown from the preted alongside sensing data and ground testing re- workspace. This process leaves cavities either spheri- sults. cal in shape or more irregular as some base rock can Grind energy index. The energy consumed by the remain attached to the spherule when it is ejected. RAT’s primary actuator while grinding a given rock, Sweeping of Fines without Grinding: The func- apart from losses such as frictional drag in the mecha- tionality of the RAT allowing it to clear cuttings can nism, is normalized to provide a representative specific be employed independently of grinding. This sweep- grind energy in terms of Joules per cubic millimeter ing operation removes unadhered or lightly adhered for each grinding operation. The calculation is per- material, exposing the rock face for observation, pri- formed over the last 0.25 millimeter of a grinding op- marily for the APXS, Microscopic Imager, Pancam, eration, where it is possible to make an accurate esti- and Mini-TES. Overlapping patterns are created to mate of the volume of rock removed. A mosaic of accommodate the larger Mini-TES field of view. Microscopic Imager images of the resulting abrasion area allows for a measurement of the area, and there- fore a projection of abraded volume. The resulting specific grind energy is considered an index of grindability.

Fig. 5: Color Pancam composite from Spirit of swept pattern on ‘Clovis’ Courtesy NASA/JPL/Cornell Arced scrape marks have been observed in some swept areas, coinciding with grinding bit and brush bit paths of motion. As of 5 January 2005, 40 sweeping operations have been performed at Gusev Crater by Fig. 4: Diagram of rock volume measurement Spirit and 8 at Meridiani Planum by Opportunity. Physical properties conclusions. An index such as

grindability describes the performance of the device References: on the given material in a given environment. De- [1] Gorevan S. P. et al. (2003) JGR, 108, 9-1 - 9-8. coupling device and environment characteristics leaves [2] Ming, D and Morris, R. (2002) Particle size analy- a bulk description of the given material’s physical sis, unpublished.