Apollo Lunar Roving Vehicle

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

Overview

The Lunar Roving Vehicle was a four-wheeled, battery-powered, crew-driven vehicle carried to the Moon on Apollo 15, 16 and 17 [8]. Boeing was prime contractor with General Motors Defense Research Laboratories responsible for the mobility system, under Marshall Space Flight Center management. It flew for the first time on Apollo 15 in July 1971 and last on Apollo 17 in December 1972 [8]. Three vehicles were flown and all three were left on the Moon.

The LRV is the only surface vehicle whose wheel design was developed against measured lunar soil simulant performance, flown, and then compared against instrumented in-situ results [2], [17], [1].

The Apollo 15 vehicle covered 27.9 km in 3 hours 8 minutes of driving, at an average speed of 9.2 km/h over the three traverses [8].

Specifications

Parameter	Value	Source
Length x width x height, deployed	3.05 x 2.13 x 1.14 m	[1], [8]
Wheelbase	2.29 m
Vehicle mass	~206 kg	[8]
Gross operational mass	~696 kg
Stowed envelope	1.78 x 1.65 x 0.94 m	[4]
Wheel diameter x width	81.3 x 22.9 cm	[2]
Wheel mass	5.48 kg
Nominal contact pressure at 254 N	4.14 kPa
Ground clearance, loaded / unloaded	35.6 / 43 cm	[8]
Minimum turn radius	3.05 m	[1], [8]
Top speed	9 to 13 km/h, terrain dependent	[1]
Estimated maximum slope capability	19 to 23 degrees	[13]
Design slope-climbing requirement	25 degrees	[2]
Range capability	120 km	[9]

Mission profile

Three vehicles flew, one each on Apollo 15, 16 and 17, and each was used on three traverses over the three surface days of its mission [8].

Parameter	Value	Source
Apollo 15 odometer distance	27.9 km	[8], [14]
Apollo 16 odometer distance	26.6 km	[14]
Apollo 17 odometer distance	34.8 km	[14]
Total distance driven	89.3 km	[14]
Driving time, Apollo 15	3 h 8 min, at 9.2 km/h average	[8]
Drive time, Apollo 17	253.5 min against a 280 min planning value	[11]
Ride time including minor stops, Apollo 15 / 16 / 17	about 180 / 199 / 269 min	[14]
Maximum range from the lunar module, Apollo 15	5.0 km	[8]
Slopes encountered, Apollo 15 / 16 / 17	0 to 12 / 0 to 18 / 0 to 20 degrees	[13]
Energy consumed, Apollo 15	52 A-h of 242 A-h available	[8]
Energy consumed, Apollo 16	98.2 A-h of 242 A-h available over 26.7 km	[10]
Navigation position error specification, Apollo 15	600 m	[14]
Navigation position error specification, Apollo 16 and 17	100 m

The distances above are traverse odometer readings, and the mission reports count differently. Apollo 16 is given as 26.7 km from extravehicular activity distances of 4.2, 11.1 and 11.4 km [10], and Apollo 17 as 36.0 km of odometer distance, of which 34.7 km was driven between traverse start and end points and the remaining 1.3 km off them [11].

The binding mission constraint was not the vehicle. Range was bounded by the walk-back limit, the distance the crew could return on backpack life support if the vehicle failed, so the 120 km range capability [9] was never approached by a 27.9 km traverse [14].

Mobility

Four wheels, each independently driven. The carcass is woven from 800 strands of 0.838 mm zinc-coated piano wire over a spun aluminum hub, with a titanium inner frame 64.8 cm in diameter acting as a bump stop; it engages at 600 N radial load and 8.26 cm of deflection, and carries a 4000 N ultimate load [2]. Chevron-shaped titanium tread strips are riveted over about 50 percent of the soil-contacting circumference [1].

The design follows from three constraints acting together. Pneumatic tires were rejected outright because a puncture is unrecoverable and rubber degrades under solar ultraviolet and vacuum; rigid rims were rejected on ride quality [2]. That leaves a metal-elastic carcass, and the woven mesh gives a large contact patch at very low mass. Nominal contact pressure under a 254 N radial load is 4.14 kPa, roughly a twentieth of a terrestrial car tire, which is what keeps sinkage small in a low-cohesion regolith [2]. The titanium inner frame exists because an unrestrained mesh would over-deflect and fail under impact: it engages at 8.26 cm deflection under 600 N and carries an ultimate 4000 N with a factor of 1.5, so the wheel has an inherent run-flat mode.

Chevron coverage was traded in lunar soil simulant. 50 percent gave the lowest power at the self-propelled point together with the highest drawbar pull at 20 percent slip, 20 percent having been fixed empirically as the slip of maximum drawbar pull without excessive power consumption and used thereafter as the ceiling for gradeability assessment [2]. Meeting the 25 degree slope-climbing requirement needs 59.5 percent slip on average, and 72.0, 79.5 and 84.5 percent slip at one, two and three standard deviations of the roughly 500 point slope-versus-slip fit across simulant conditions LSS1 to LSS4; wheel travel over a 0.5 km traverse on a 25 degree slope could therefore reach 3.2 km. Slopes encountered on Apollo 15 were no more than 12 degrees [2].

Candidate wheels were compared in loose dry sand before selection. Slope-climbing capability was 28 to 30 degrees for the Bendix wheel and 15 to 20 degrees for the Boeing-GM and Grumman wheels, none reaching 35 degrees; adding aggressive grousers raised Bendix and Grumman above 30 degrees and the modified Boeing-GM wheel to about 25 degrees [17]. Level-surface power under a 220 N load was 4 W.h/km for Bendix, 6 W.h/km for Boeing-GM and 10 W.h/km for Grumman. Single-wheel test predictions of vehicle slope-climbing ability run about 1 to 2 degrees of slope conservative [17].

Suspension is a pair of parallel triangular arms per wheel reacting into a torsion bar, with a velocity-square damper between chassis and upper arm [1]. Combined suspension and tire deflection gives 35.6 cm of ground clearance loaded and 43 cm unloaded [8]. Steering is double Ackermann, front and rear pairs driven by independent electric rack-and-pinion assemblies, 22 degrees outer and 53 degrees inner, lock to lock in about 5.5 s, minimum turn radius 3.05 m. The two assemblies are electrically and mechanically independent and either can be disengaged and its wheels centred and locked, so a single steering failure degrades rather than ends mobility. That redundancy was exercised [8]: front steering was inoperative for the whole of the first Apollo 15 traverse and the crew drove on rear steering alone, with the consequence that the rear wheels slid into features the driver was steering to avoid, and the vehicle rotated through 180 degrees at least once. Cycling the forward steering circuit breaker restored it before the second traverse.

Terramechanics

The soil-vehicle model used for LRV prediction is the Bekker Land Locomotion Laboratory model C, for flexible-tire wheels on soft soil [1]. The parameter set recommended in the NASA request for proposals, and used for the Boeing Soil A, B and C models, was:

Parameter	Value
Cohesive modulus of deformation k_c	0 to 0.70 kN/m^(n+1)
Frictional modulus k_phi	81.4 kN/m^(n+2)
Deformation exponent n	1.0
Shear deformation modulus K	1.8 +/- 0.8 cm
Cohesion c	0 to 0.34 kPa
Internal friction angle phi	35 +/- 4 degrees
Bulk density	0.80 to 1.60 g/cm3
Wheel-to-surface friction coefficient	0.6

Values from [1]. Bekker parameters measured directly on lunar soil simulant at the Waterways Experiment Station gave k_c of 0.07 to 0.16 (kN/m)(cm^-n), k_phi of 58.19 to 74.63 (kN/m2)(cm^-n) and n averaging 0.49 to 0.51 [17], which differ from the design set by more than an order of magnitude in k_phi and by a factor of two in n.

Wheel performance was characterized at the US Army Engineer Waterways Experiment Station in a ground-basalt lunar soil simulant from Napa, California, prepared at graded densities LSS1 through LSS5 [1], [3]. Results are reported dimensionlessly: pull coefficient P/W, torque coefficient M/Wr, power number PN, against slip and against penetration resistance gradient G.

Three findings from that campaign matter beyond Apollo. First, in the simulant, pull, torque and power number were effectively independent of wheel speed from 0.44 to 3.14 m/s [3]. This is not obvious, and the report attributes the constancy to two effects canceling: dynamic soil momentum raising pull with speed, and air pore pressure in the low-permeability simulant suppressing it. Since there is no pore air on the Moon, the authors concluded LRV wheels could be more efficient at speed in flight than in the laboratory. Second, performance scaled with soil strength as expected: at a pull coefficient of 0.25, slip was 8 percent in the denser LSS5 against 11.5 percent in LSS4, and maximum sustainable slope without excessive power was 23 plus or minus 5 degrees in LSS5 against 19 plus or minus 6 degrees in LSS4 [3]. Third, chevron direction was immaterial: running the wheel backwards, with the chevrons reversed, produced no measurable performance difference.

Flight results were better than the terrestrial predictions. Apollo 15 wheel sinkage averaged 1.25 cm at the 2.3 percent slip used to calibrate the odometer, ranged from imperceptible to 5 to 7 cm, and reached about 13 cm only once, at high slip at Station 8 [1]. Median computed wheel slip on level ground was 2.1 percent. Measured traction drive energy consumption was 24.8, 30.3 and 13.6 W-h/km for the three EVAs, and the crew reported that traction was excellent uphill, downhill and under acceleration, with 10 km/h reached in about three vehicle lengths and very little slip [8]. Predicted energy consumption overestimated measured values by about 30 percent at the median, because the flown soil was firmer and the resistance lower than the design case. The margin came from conservatism in the simulant work.

Across the three missions the vehicle worked progressively harder terrain: slopes of 0 to 12 degrees on Apollo 15, 0 to 18 degrees on Apollo 16 and 0 to 20 degrees on Apollo 17, against an estimated maximum capability of 19 to 23 degrees [13]. The soil mechanics experiment read the tracks as data: average sinkage 1.25 cm, constant slip of only 2.3 percent, sharp chevron imprints and excellent flotation throughout, with relative density inferred from vehicle tracks averaging 62 to 71 percent and a penetration resistance gradient of 1 to 6.6 N/cm3, averaging 3.8.

The Surveyor 3 and 7 surface samplers had produced a lunar soil model of a slightly cohesive sand at about 1.5 g/cm3 bulk density, 0.7 kN/m2 cohesion and 35 to 37 degrees friction angle [13]. Those parameters, carried forward as Model B, reproduced the measured Apollo 16 and 17 LRV energy consumption. A robotic scoop on a static lander, five years before the first drive, had characterized the soil well enough to predict a crewed vehicle’s power draw.

The mission had one soft-soil arrest: a single occurrence of wheel spin in loose soil at the ALSEP site was resolved by the crew lifting the vehicle bodily to fresh ground [8]. A robotic vehicle has no equivalent recovery, which is the reason later designs, VIPER among them, carry a non-rotational extraction mode.

Power and energy

Two non-rechargeable silver-zinc batteries, each nominally 36 V and 120 A-h, giving 240 A-h installed and about 64 kg [8], [9]. Both are normally discharged in parallel at approximately equal load, and the circuit is arranged so that the whole load can be transferred to one battery if the other fails. Each is protected by a relief valve opening at 3.1 to 7 psi differential [8].

Energy was never the limiting resource. Apollo 15 consumed about 52 A-h of the 242 A-h available, leaving 190 A-h, at an actual rate of 1.87 A-h/km against a predicted 3.67 A-h/km [8]. Apollo 16 consumed 98.2 A-h of the 242 A-h available over 26.7 km [10], and Apollo 17 reported energy rates between 1.54 and 1.8 A-h/km [11], consistent with Apollo 15. The real constraint was crew consumables and the walk-back limit: the vehicle could never be driven further from the lander than the crew could walk home on backpack life support.

Regenerative braking was evaluated and rejected. With the series brush motors flown, generator behavior was not predictable enough to rely on, so conventional mechanical drum brakes were fitted to every wheel; analysis of the brushless alternative showed regeneration would have returned less than 2 percent of total energy in any case [9].

Thermal

The LRV was designed for lunar morning operation only, which narrows the thermal problem from the full cycle range of about plus or minus 120 C to a 0 to 100 C band and a service life of about 78 hours [9]. There are no radioisotope heaters and no active loop.

Thermal control is entirely passive and duty-cycled: special surface finishes, multilayer insulation, space radiators, second-surface mirrors, thermal straps and fusible-mass heat sinks. The operating concept is to store energy in the fusible masses while driving and reject it to deep space while the vehicle is parked between sorties, with radiators on top of the signal processing unit, the drive control electronics and both batteries [8]. The radiators carry dust covers that the crew opened at the end of each traverse and that closed automatically once the unit had cooled.

This is the design’s known weakness, and Apollo 15 demonstrated it. Battery 2 had not cooled enough by the start of the second EVA for its cover to close, attributed to differing dust accumulation on the two thermal mirrors; by the third EVA little cool-down had occurred on either because of further dust [8]. Battery temperatures rose from 68 and 78 F at second-EVA activation to 108 and 113 F at the end of the third, still within limits. The crew judgment recorded in the mission report is the durable lesson for any passively cooled lunar surface system: the covers did not seal tightly enough against their Velcro edges to keep dust off the mirrors, and only a small amount of dust on the surface will preclude the desired cool-down. Drive motors carried thermal instrumentation with a warning flag at 204 C [1].

The problem recurred and worsened on the later missions. On Apollo 16 the crew brushed the radiator surfaces and opened the battery covers after the first EVA, the covers then failed to close before the second, and battery 2 reached 143 F by the close of the third [10]. On Apollo 17 initial power-up temperatures were 95 and 110 F against a predicted 80 F each, and battery 2 temperature indication went off-scale low at the start of the third EVA [11]. The dominant contributor was mechanical rather than thermal: the right rear fender extension was knocked off during the second Apollo 16 EVA and again before the first Apollo 17 traverse, after which the unshielded rear wheel threw dust over the crew, the console and the communications equipment. Dust control and passive thermal control are the same problem on a wheeled lunar vehicle, and losing the fender extension was enough to couple them.

Compute and avionics

The LRV carried no general-purpose computer. Its avionics are analog and electromechanical throughout: hand controller potentiometers feeding pulse-width modulators in the drive control electronics, a current-level detector inhibiting drive above about 25 A, and a signal processing unit performing the dead-reckoning computation [9]. Motor current is limited electronically to 22 A. Radiation tolerance was not a design driver, because the surface mission is measured in days rather than years and the electronics are discrete rather than microprocessor-based.

Autonomy

None. The LRV has no autonomous function of any kind; it is driven directly by a crew member with a T-handle controller, forward speed proportional to forward deflection past a 1.5 degree deadband, braking by pivoting the handle rearward, and a required full stop before any direction reversal [8].

Navigation is the one function performed onboard without the driver. It is dead reckoning: a sun-aligned directional gyro supplies heading, odometer pickups on each traction drive supply distance at nine pulses per wheel revolution, and the signal processing unit integrates these into displayed range and bearing back to the lunar module [1], [8]. Gyro alignment is a ground-in-the-loop procedure: the crew relay pitch, roll and sun angle to Earth, an initial heading is computed there, and the crew slew the gyro with a torquing switch until the heading indicator matches. Odometry is calibrated against an assumed constant 2.3 percent wheel slip bias [8].

The mechanism is a slip-rejection scheme implemented in hardware. Each traction drive produces nine pulses per wheel revolution, 0.245 m of travel per pulse [14]. Rather than averaging the four wheels, four divide-by-three counters are arranged so that the third fastest wheel drives the distance increment, in steps of 0.735 m. The slowest wheel is discarded because it may be dragging, and the fastest because it may be spinning free, so a single wheel losing traction contributes nothing to the distance estimate. The signal processing unit accumulates north and east components as the products of the increment with the cosine and sine of heading, then converts the Cartesian accumulation to range and bearing using a CORDIC algorithm.

Heading comes from a Lear Siegler directional gyro with drift below 5 deg/h in the laboratory and below 10 deg/h expected on the Moon, torqued to the ground-computed true heading three minutes after power-on, with a constant torquer voltage compensating Earth rotation [14]. Displays resolve heading and bearing to 1 degree and range and total distance to 0.1 km, with specified accuracies of plus or minus 4.5 degrees in heading, 4.6 degrees in bearing, 420 m in range at 5 km, and 1 percent in distance.

Apollo 15 showed essentially no gyro drift, closure errors on the three traverses of 0, 100 and 200 m, a position error of about 0.3 km at the maximum 5.0 km range, and no traverse realignments required [8]. Apollo 16 and 17 were better still: maximum position error of 100 m on every traverse of both missions, closure of 0.0 m on the first Apollo 16 and first Apollo 17 traverses, and all traverses inside a 100 m specification against the 600 m applied to Apollo 15 [14].

Hazard detection was entirely visual and the mission report quantifies its limits: forward visibility was excellent except when driving toward the zero-phase direction, where contrast vanishes, and 1 m craters were not detectable until the front wheels were within 2 to 3 m [8]. Route selection was accordingly by crater avoidance rather than crater traversal.

Communications

The LRV carried the Lunar Communications Relay Unit and the ground-commanded television assembly, installed on the forward chassis by the crew after deployment [8]. The LCRU provided a direct-to-Earth link independent of the lunar module, relaying crew voice and telemetry and transmitting color television, and receiving voice and camera pointing commands from Earth. It carries an S-band high-gain antenna mounted on the forward chassis and an S-band low-gain antenna in the left inboard handhold, with redundant S-band transmitters and receivers and a 70 kHz command subcarrier to the television assembly [8]. The unit runs on its own batteries with a backup connection to the LRV power system. Its thermal design mirrors the vehicle’s: thermal blankets, second-surface radiating mirrors, and change-of-phase wax packages to absorb peak heat.

The television camera was pointed from Earth through the LCRU command link [8], making a flight controller the effective camera operator. Ground remote control of a lunar surface vehicle had been rehearsed before flight in NASA MSFC tests of remote driving and navigation [16]. Published downlink rate figures for the LCRU are not in the sources held here.

Payload and instruments

The LRV is a carrier rather than an instrument platform. Its cargo on a given traverse comprised the lunar hand tool carrier, the Apollo lunar surface drill, sample containers, the 16 mm and 70 mm cameras and the 500 mm lens, the color television camera and its control unit, and on Apollo 15 the laser ranging retroreflector [8]. The vehicle’s own instrumentation is limited to engineering telemetry: battery voltage, current, remaining ampere-hours and temperature, motor temperatures, and the navigation displays for pitch, roll, speed, heading, total distance, and range and bearing to the lunar module. Displays use a promethium radioluminescent material so they remain readable in shadow.

The vehicle is itself a soil mechanics instrument, and was used as one. Wheel track depth, tread imprint sharpness and observed slip were read as measurements of soil strength, and cone penetrometer readings were taken inside and beside the tracks: 5.97 lb/in2 per inch of gradient in the upper 2 cm of an LRV track against 2.98 adjacent to it [1], a direct measurement of compaction by a single wheel pass.

Modes of operation

The vehicle has no software modes. Its operational states are configurations selected by switch: four independent DRIVE ENABLE switches allowing any wheel to be dropped from the drive train, a decoupling mechanism permitting any wheel to free-wheel mechanically, a choice of either of two pulse-width modulators to drive any wheel, and selection of either power bus from either battery for each load [8], [9]. Steering has three usable modes, double Ackermann, front only and rear only, the latter two being failure configurations that were both flown on Apollo 15.

Apollo 15 crew evaluation of those steering modes was a result rather than a specification: rear-only steering caused the front wheels to dig in and the rear to break out under large corrections; front-only was tried and abandoned because the unpowered rear wheels wandered; simultaneous four-wheel steering was more sensitive than desired but was used for the remainder of the mission as the best available [8].

The parked state is the thermal recovery state described above, and is the only mode with a duration constraint attached to it.

Ground operations

Driving was performed by the crew in real time with no ground involvement, so ground operations reduce to three functions. Traverse planning was done pre-mission and revised in the flight control room against consumables and the walk-back limit. Gyro alignment required a ground computation of initial heading from crew-relayed pitch, roll and sun angle [8]. The television camera was pointed from Earth through the LCRU command link, which made a flight controller the effective camera operator throughout each EVA.

Deployment was a crew procedure. The vehicle was stowed folded in lunar module quadrant 1 at 1.78 m wide, 1.65 m high and 0.94 m deep, having been folded to 50 percent of its length and 70 percent of its width, with forward and aft chassis sections folded back over the center section and the wheel suspensions rotated inward [4]. Deployment is gravity-driven and rate-limited by hand: the crew pull a D-ring to release the attachment pins, a pushoff spring starts the vehicle rotating about lower hardpoints, and descent rate is controlled by an astronaut paying out a deployment tape over braked reels using worm-gear cable drums with self-locking gears. Torsion springs unfold the forward chassis and torsion bars the aft chassis and wheels; spring-loaded shear pins, two in each of four chassis hinges, latch the chassis automatically once unfolded, and the wheel struts release by cable once their chassis is within 10 degrees of fully unfolded [4]. The requirement was deployment in thermal vacuum within 15 minutes, by crew in pressure suits, with the lunar module tilted up to 14.5 degrees on up to a 6 degree slope [4]. On Apollo 15 the only deployment problem was the saddle fitting binding under preload from lunar module tilt, closed by crew training rather than hardware change [8].

Technologies developed

The wheel is the durable output. The wire-mesh carcass with a rigid inner bump stop and partial chevron tread, sized against measured simulant performance at a target slip, is the canonical low-mass planetary wheel and remains the reference design against which flexible metal wheels are argued [2]. The development chain is documented from SLRV through MOLAB to the LSSM concept trade, where eight wheel concepts were scored and the wire mesh and hoop spring taken forward, through the fatigue work that established piano wire as the least prone to vacuum welding of eight candidates and the endurance tests that carried a mesh wheel to 85,000 cycles.

Second, the terramechanics dataset itself. The WES simulant campaign established a dimensionless framework, pull coefficient, torque coefficient and power number against slip and penetration resistance gradient, that is still how rover wheel performance is reported [3]. The flight comparison, where the terrestrial prediction overestimated energy consumption by about 30 percent [1], is one of very few opportunities the field has had to check a soft-soil model against a real extraterrestrial traverse.

Third, the sealed traction drive: a series brush motor mated to an 80:1 harmonic drive, hermetically sealed and running in dry nitrogen at about 5.17 N/cm2 with Krytox 143AZ lubricant, with the gas serving both to lubricate and to conduct heat out of the gearbox [9]. Sealing a gas atmosphere around a geartrain rather than attempting dry vacuum operation is a pattern reused widely since.

Fourth, the dead-reckoning navigator, and specifically the third-fastest-wheel odometry rule, which is the earliest flown example of rejecting slipping and dragging wheels in hardware rather than modeling them, and delivered 100 m position accuracy over 30 km traverses with no external reference at all [14].

Fifth, the folding and gravity-deployed stowage scheme [4], which remains the reference for fitting a full-size surface vehicle into a lander envelope.

The negative results carried forward as well. Dust on second-surface mirrors defeats a passive radiator [8]. Velcro edge seals do not exclude lunar dust. Visual hazard detection saturates at a few meters in the zero-phase direction.

References