Scarab

NASA. Public domain (NASA / US government work).

Scarab is a 280 kg four-wheeled prospecting rover built at Carnegie Mellon University in 2007 to carry a 1 m coring drill into permanently shadowed lunar polar craters [1], [12]. It never left Earth.

The problem

A cold-trap prospector has to find and sample water ice and other volatiles that survive only because they are cold. The design environment for Scarab was hard vacuum, a 40 K ground and a 3 K sky, with no sunlight for power or cameras [1]. Survey compilations list the same envelope of temperature extremes, dust and regolith bearing limits for every lunar rover concept, and most of the concepts in them were design targets that were never built [11].

Three requirements follow. The vehicle must drive across unconsolidated slopes on crater walls. It must hold a stiff base for drilling, because a coring drill thrusting at 100 to 200 N or more needs weight on the ground to react it, and that requirement set Scarab’s mass [1]. It must do both without ambient light, so navigation uses active laser sensing [10].

The ground it works on is characterized poorly. The geotechnical values in the standard engineering reference come from Apollo and Luna sites near the equator [14], and the polar regolith that Scarab was built for has not been sampled in place [10].

What has flown

No Scarab flew, and no rover has drilled lunar polar regolith. The flown items that bear on the design are few.

Honeybee Robotics’ TRIDENT drill, a 1 m rotary-percussive design with a twenty-year lineage that includes Resource Prospector, was built in two identical flight units, one for the PRIME-1 payload of the IM-2 lander and one for VIPER [28]. VIPER was canceled before launch, so only the PRIME-1 unit was in a position to reach the Moon.

LEV-1, a JAXA hopping rover, operated on the Moon after the SLIM landing in January 2024 [22]. Its single wheel carries a 90-degree-tilted octagonal grouser chosen with a resistive-force-theory model and checked in sand and discrete element simulation before flight [22]. It is a different vehicle class from Scarab and the only flown grouser result among the works cited here.

Scarab itself is a ground prototype. The 2008 design paper reports a continuous 1 km traverse, a 50 cm trench bridge and rock obstacles of 40 to 50 cm climbed, all at Earth gravity on rubber skid-loader tires [1]. The 2009 field paper reports the same vehicle with the NORCAT coring drill and the RESOLVE volatile payload at the NASA Glenn SLOPE facility, Moses Lake dunes and the Mauna Kea analog site, with more than 3 km of autonomous laser-based traverses [10].

What is being tried now

Reconfigurable chassis and the Resource Prospector lineage

Scarab’s active ride height and wheelbase adjustment sits on a passive differenced rocker. Leaning into a 20 to 25 degree slag slope halved downhill slip from 15 percent to 7 percent over single 11 m runs [1]. Instrumented tests on GRC-1 simulant reported a cut from 37 percent to 15 percent at 20 degrees [10]. Status: ground tested only.

Resource Prospector carried the idea into a 300 kg solar vehicle with offset four-wheel steering and active suspension, so the array could stay sun-pointed while driving [4]. The prototype was built in 2014 and 2015, and the mission was canceled in 2018 [4], [5]. Its launch lock was a four-post design released by shape memory alloy actuators, a TRL 4 proof of concept with no quantitative release data [8].

VIPER was the next vehicle, a 450 kg rover with three mobility actuators (suspension, steering, drive) packaged into corner modules [11]. Its dust protection is a 42-layer outer barrier over each module with graded labyrinth, felt and PTFE seals [19]. VIPER was also to use torque transducers in its suspension joints and a quadrature encoder pattern on the wheel rim to infer wheel normal force and sinkage, and to use the suspension as a geotechnical instrument and the IMU as a gravimeter and seismometer [23], [15]. The mission specification is in the proposal information package [2]. Status: flight hardware built and ground tested, not flown, so none of these measurements were taken on the Moon.

Drilling and sample integrity in cold regolith

A one meter class auger drilled into cryogenic NU-LHT-3M simulant doped with 5 weight percent water produced four sealed samples out of twelve attempts, and those four lost on average 30 percent of their water [7]. Bringing the simulant bed to -130 C took about 48 hours [6]. A mass spectrometer viewing the cuttings pile distinguished dry, 2.5 and 5 weight percent layers while TRIDENT drilled four holes at 1.6e-6 Torr [1]. In Antarctic field drilling of fifteen boreholes with the TRIDENT engineering unit, telemetry alone identified an ice-rock transition near 62 cm [3]. Status: ground tested in simulant or terrestrial ice.

Excavation without a thrust axis

Excavators avoid the reaction problem when their loads cancel. The IPEx excavator qualified one actuator design for wheel, arm and bucket drum roles to TRL 5 [24], with the actuator-level qualification in a dynamometer campaign at KSC [4]. A 3D-printed bucket drum for the MoonBot modular rover measured 777.54 kg/h of continuous excavation in silica sand [13]. A coring drill pushes along one axis, so the same cancellation does not apply to Scarab’s task. Status: ground tested.

Terramechanics and reduced-gravity prediction

Single-wheel tests on a small skid-steer rover found that 45 degree diagonal grousers reduce point-turn sinkage, at one slip level [3]. Two parabolic flight campaigns showed that matching only wheel load in a 1 g test overstates mobility, and graded granular scaling laws against that data [7]. A scaling law enhanced with slope terms was validated against a discrete element model only [20], and a later DEM comparison of scaling methods used 1/6 g simulation as its reference [21]. A discrete element model of a Rashid-1 style wheel, calibrated only against angle of repose, predicted nearly unchanged traction at lunar gravity, with no reduced-gravity experiment behind it [16]. Traction at speeds up to 1.0 m/s depends on speed in a way low-speed tests miss [25]. A continuum simulation of the six-wheel Yutu rover under lunar gravity reproduces DEM single-wheel curves at lower cost, and has no flight telemetry behind it [30]. Status: ground tests and simulation.

Simulants and test beds

GRC-3b-DST adds real cohesion, 11.5 kPa ultimate at 81 percent relative density, to a Glenn simulant that had none [12]. A drop-tower shear cell obtains strength at small-body and lunar acceleration [8]. The case for VIPER after cancellation is made in terms of instruments and risk reduction, not results [15].

The numbers

Scarab as built [1]:

Parameter	Value
Mass	280 kg
Weight, Earth surface	2740 N
Weight, lunar surface	450 N
Footprint	roughly 1.5 by 1.5 m
Drill tower height, upright	2.2 m high stance, 1.6 m low stance
Average power available	100 W
Nominal power	175 to 200 W
Idle power	24 W
Locomotion speed	3 to 6 cm/s
Wheel diameter	65 cm
Track width	1.4 m
Wheelbase	0.8 to 1.4 m, 1.2 m nominal
Center of gravity height	0.48 m low, 0.64 m nominal, 0.74 m high stance
Static pitch-over angle	56 degrees low, 43 degrees nominal, 30 degrees high stance
Static roll-over angle	61 degrees low, 53 degrees nominal, 49 degrees high stance
Maximum straddle	55 cm
Approach and departure angle	105 degrees nominal stance
Breakover angle	115 degrees nominal stance

Tip-over angles were measured on a tilt table [1]. Drawbar pull reached 2000 N in mixed-grain sand, about 0.7 times vehicle weight [1].

Resource Prospector parameters [4]:

Parameter	Value
Mass, including payload	300 kg
Dimensions	1.4 x 1.4 x 2 m, mast included
Speed made good	0.5 cm/s

The launch lock allocation was 170 kg rover and up to 105 kg payload, with net launch mass not above 325 kg [8]. The Virtual Bumper laser-dot hazard detector claimed a 98 percent success rate with no sample size reported [4].

What is not established

All Scarab mobility and drilling tests ran at Earth gravity on rubber skid-loader tires. Drawbar pull, slip and obstacle climbing were not characterized in reduced gravity, and gravity offload was only planned [1], [10]. The slip figures are single 11 m runs [1], and the two results come from different soils and slopes. Drawbar pull is published as a fraction of vehicle weight, without load-slip curves or soil parameters.

The 100 W average is a supply assumption from an ASRG electrical simulator, the thermal design for a 40 K ground is a first-order analysis, navigation in darkness was future work, and drilling loads in the tests were described as relatively moderate [1].

No drill has operated in lunar polar regolith, and the water-loss data rest on four usable samples [7]. Reduced-gravity predictions are validated against short parabolic flights or against other models [7], [20]. Polar regolith properties are inferred from equatorial data [14], and Resource Prospector and VIPER were canceled before measuring them [4], [23]. The Virtual Bumper figure has no stated test population [4], and the Scarab web page gives no numbers at all [12].