Scratching the Moon

A BOUNCY LANDING

The 6-month hiatus after Surveyor 2 was not due to concern over the loss of that spacecraft, but to the wait for the restartable version of the Centaur. The two-burn configuration traded payload capacity against the hardware to restart the engines and the cryo-propellants that would be vented while coasting in parking orbit, but it offered Surveyor launches in winter months,[33] considerably lengthened the launch windows, and increased the flexibility in selecting the arrival time for optimal illumination at the landing site. The Centaur stage demonstrated its restart capability by a launch on 26 October 1966, thereby completing its test program.

The target for Surveyor 3 was a 60-km-diameter circle in the southeastern part of Oceanus Procellarum, centred 120 km southeast of the crater Lansberg. Its attraction was that although it was crossed by a ray from Copernicus 370 km to the north, at telescopic resolution it was sparsely cratered. The smooth patch of mare material was broken about 20 km to the west by rough hummocky terrain and isolated hills, and was bounded to the east by low ridges. It had been photographed at medium resolution by Lunar Orbiter 1 as target I-P-7, and at high resolution by Lunar Orbiter 3 as III-P-9c. These pictures revealed the presence of sub-telescopic craters in the target circle – with one, just over 1 km in diameter, situated very near the aim point. As a smooth-looking patch of Oceanus Procellarum, it bore a similarity to the Surveyor 1 landing site in the Flamsteed Ring, some 650 km to the west.

Surveyor 3 lifted off from Pad 36B at 07:05:01 GMT on 17 April 1967. The Atlas jettisoned its booster section at T+ 142 seconds and the sustainer engine shut down at T + 238. Once free of the Atlas, the Centaur established the desired circular parking orbit at an altitude of 160 km – with insertion at T+569 seconds. The coasting phase would vary between 4 minutes and 25 minutes, depending on the geometry of the translunar injection – in this case it was to be 22 minutes 9 seconds. While coasting, the Centaur first fired two 50-pound-thrust hydrogen peroxide thrusters to settle the remaining propellants in their tanks, then continuously fired two 3-pound thrusters to maintain this condition. It had two clusters of 3.5 and 6- pound thrusters to control its attitude, and while maintaining its longitudinal axis to the local horizontal it rolled at a rate of 0.17 degree per second in order to even out solar heating of its payload and vented any propellant boil-off.

Owing to the predawn launch, the Centaur emerged from the Earth’s shadow at 07:21:25. About 40 seconds prior to the translunar injection, the 50-pound thrusters fired again to guarantee that the propellants would enter their feed pipes. The main engines were shut down when the inertial guidance system sensed that the requisite velocity had been achieved – in this case at 07:38:49, after a 108-second burn. As on earlier missions, after it had configured and released the spacecraft, the spent stage performed the separation manoeuvre. Once free, Surveyor 3 stabilised itself and then adopted its cruise attitude. The midcourse manoeuvre at 05:00:03 on 18 April lasted 4.3 seconds and the 13.8-ft/sec change in velocity was entirely devoted to achieving the ‘critical component’ required to reduce the divergence from the centre of the target circle from the initial 480 km down to a mere 5 km.

The pre-retro manoeuvre in which the spacecraft departed from its cruise attitude involved starting a yaw of-158 degrees at 23:23:30 on 19 April and a pitch of-76.8 degrees at 23:30:17 to align the thrust axis with the velocity vector. The final roll of -64 degrees initiated at 23:34:35 was to optimise the RADVS. The initial approach was at 23.6 degrees to the local vertical. This would require a significant gravity turn during the vernier phase of the descent to force the trajectory to vertical.

The altitude marking radar was enabled at 23:59:33, and issued its 100-km slant – range mark at 00:01:12.829 on 20 April. The delay to the initiation of the braking manoeuvre was specified as 5.090 seconds. The verniers ignited precisely on time, and the new ‘high-impulse’ retro-rocket 1.1 seconds later – at which time the vehicle was travelling at 8,618 ft/sec. The acceleration switch noted the peak thrust of 9,550 pounds fall to 3,500 pounds at 00:02:00.587, giving a burn duration of 40.0 seconds. After allowing the thrust to tail off, the casing was jettisoned at 00:02:12.429. At burnout, the angle between the vehicle’s thrust vector and velocity vector was 21.1 degrees.

When the RADVS-controlled phase of the flight began at 00:02:14.642, the slant range was 36,158 feet (and because the velocity vector at burnout was offset to vertical, the altitude was 32,900 feet) and the total velocity was 483 ft/sec (and since the vehicle had maintained its thrust along the velocity vector extant at the time of retro ignition, the longitudinal rate was 462 ft/sec). The altimeter had locked on at a slant range of 43,700 feet, only to drop out again. So when the RADVS was given control it aligned the thrust axis along the velocity vector extant at retro burnout

The descent of the Surveyor 3 spacecraft depicted in two sections, one for slant ranges above 1,000 feet and the other below 1,000 feet.

and flew with the verniers at 0.9 lunar gravity, very slowly accelerating as it descended. When the altimeter locked on again at 00:02:15.786, attitude control was switched from inertial to radar, and the thrust axis was swung in line with the instantaneous velocity vector to initiate the gravity turn. On intercepting the ‘descent contour’ at 00:02:33.816, the slant range was 22,300 feet and the speed was 495 ft/sec. By the 1,000-foot mark at 00:03:53.023, the vehicle was descending almost vertically with a sink rate of 103.3 ft/sec. When the 10-ft/sec mark was issued at a height of 46 feet at 00:04:10:623, it seemed to be home and dry.

But at 00:04:13.275, at a height of 30 feet, one of the three angled radar beams lost its lock on the surface. As the flow of data to the closed-loop computer abruptly ceased at 00:04:13.387, the control system reverted to its inertial guidance system to maintain its attitude and throttled the verniers to cancel out 0.9 of lunar gravity. But because the RADVS was no longer operative, it was unable to issue the 14-foot mark intended to cut off the verniers!

At 00:04:18.050 the vehicle touched down with a vertical rate of about 6 ft/sec. Although it was level at this time, the ground was sloping down to the west, causing leg no. 2 to make contact first. In response to the tilt induced by the other two legs touching down, the flight control system – which was in attitude-hold mode and did not realise that it was on the ground – increased the thrust of

The axial forces on the shock absorbers of the three landing legs of Surveyor 3 from first touchdown to finally coming to rest.

verniers no. 1 and 3 to re-establish a level attitude, and this additional thrust caused the vehicle to lift off.

After peaking at about 38 feet, the vehicle made second contact at 00:04:42.030 some 50 feet west of the initial point, this time at a vertical rate of 4 ft/sec. Just as previously, the slope caused leg no. 2 to touch down first and in its effort to hold its attitude the vehicle lifted off again. The engines were cut off by a command from Earth at 00:04:53.907 – at which time the vehicle had peaked at a height of 11 feet and was at 3 feet and falling. Since a portion of the thrust had been aimed laterally at each liftoff, this had built up a horizontal component, with the result that when the vehicle struck the surface its vertical rate was only 1.5 ft/sec but it had a horizontal rate of 3 ft/sec. The elasticity in its legs caused it to rebound several inches and hop another 18 inches further downslope before it settled at 00:04:54.420, some 36 feet west of its second point of contact. The gyroscopes indicated the lander to be tilted towards the west at an angle of about 12.5 degrees from vertical.

An investigation concluded that the most likely cause of the RADVS dropping out as it neared the surface was that its logic ordered a ‘break lock’ as one of the beams crossed a field of rocks – to a microwave radar, angular rocks would have appeared much as broken mirror fragments would to a searchlight. The circuitry was designed to make the radar tracking circuits select the strongest signal if several were present. It was essential to ignore antenna ‘side lobes’ when the radar was preparing for the gravity turn. As Surveyor 3 made the final vertical descent, the scintillating side lobe had obliged the system’s logic to break its lock. As the probability of losing lock on the main beam during the vertical phase of the descent was negligible, it was decided that this problem would be eliminated on future missions by having the flight control system inhibit this side-lobe rejection logic upon receiving the 1,000- foot mark.

The first 200-line picture was received 58 minutes after landing, and a total of 55 wide-angle pictures were obtained in this mode. For this mission, a small visor had been added to the hood of the camera to prevent direct sunlight from penetrating the optical train in the hope of reducing the glare that had afflicted Surveyor 1 when the Sun was above 45 degrees of elevation. Surveyor 3 arrived approximately 23 hours after local sunrise, and in the orientation in which the vehicle landed the camera was on the eastern side with the Sun 11 degrees above the horizon. There was therefore surprise that many of the preliminary pictures were partially or completely obscured by a veiling glare. It was concluded that either engine efflux or fine particles stirred up by the engines during the ‘hot’ landings had coated part of the camera’s mirror such that when that part of the mirror was directly illuminated by sunlight the view of the lunar surface was obscured. In addition, any scene that included terrain which strongly reflected sunlight was similarly degraded. Later, intermittent sticking of the mirror in both its azimuth and elevation motions implied that dust had penetrated its mechanism. The hood rotated in azimuth with the mirror, and the mirror could be rotated in elevation to seal the hood, but the engineers had been reluctant to start off in that configuration in case the mirror failed to open. A better hood was already in development. As events would show, the camera’s operational issues would impair the imaging schedule and the glare would degrade the results. A telemetry problem meant that scanning for the Sun and Earth could not start until 06:32. This issue had appeared at the time of the second contact in the protracted arrival. It proved to be a signal processing failure. The fact that the inoperative RADVS lost its high-voltage supply at the same time implied that the signal processing problem was the result of an electrical arc. After a detailed study of the performance of the system identified a number of short circuits, a work-around was devised to minimise the impact on the surface activities. Meanwhile, the Sun and Earth acquisition was completed at 08:15, and the first 600-line picture was taken at 08:42. By handing over in succession, the Deep Space Network stations at Goldstone in California, Canberra in Australia and Madrid in Spain maintained continuous communication with the lander.

An analysis of the early pictures determined that the horizon was 5 degrees higher than it would have been if the lander were on a level plain – indicating that it was in a shallow depression. From the tilt, it was inferred to be on the eastern interior wall of a medium-sized crater.

In-flight tracking could locate the landing site only to within a few kilometres, but the crater in which the vehicle had settled was able to be identified by comparing the landscape observed at ground level with the overhead view of frame H-154 taken by Lunar Orbiter 3 – although obviously the lander was not present at the time that this picture was taken. This showed that Surveyor 3 was within 2.8 km of the aim point. The area was in frame H-125 taken by Lunar Orbiter 4 in May 1967, but because that mission was mapping from high altitude the resolution was insufficient to show the lander – nevertheless, the resulting refinement of the selenodetic grid enabled the coordinates of the site to be measured to an accuracy of better than +0.01 degree in each ordinate.

Once the crater in which Surveyor 3 landed had been found in overhead imagery, its diameter was measured at about 200 metres. It was actually the largest of a tight cluster of craters arranged in a pattern which would later be dubbed the Snowman. Photoclinometry of H-154 suggested that the crater was about 20 metres deep, that there was a smooth transition from the concave floor, that the slopes of the interior walls averaged 10 to 15 degrees, and that the rim was low and gently convex. There was an inflection in the profile from concave to convex about half way between the centre and the rim crest, in both radial and vertical directions. The ‘ground truth’ of Surveyor 3 offered a means of checking the automated photoclinometry of overhead imagery – in particular, the depth of 20 metres was seen to have been overestimated by about 5 metres.

The overhead view resolved about 100 small craters scattered over the floor, inner slopes and rim of the main crater. These ranged in size from 25 metres down to the effective limiting resolution of 1 metre. Most had gentle interior slopes and rounded rims, but a few had steep interior slopes and sharp rims. Since blocks were of higher albedo than the surface material it was possible to discern blocks down to half a metre in size, and it was evident that most were related to three of the largest craters superimposed on the main crater. By taking bearings on features and relating these to the overhead perspective, the location of the lander could be pin-pointed to within 0.5 metre – it was almost half way between the centre and the rim crest. Because it was at the inflection of the slope, its foot pads were about 7 metres below the rim

A photograph by the 61-inch reflector of the Lunar and Planetary Laboratory of the University of Arizona showing the part of Oceanus Procellarum to the southeast of Lansberg (top left corner) to which Surveyor 3 was assigned. Note the hummocky terrain to the west and the wrinkle ridges to the east. The outline shows the area covered by the next illustration.

crest and about 7 metres above the centre of the cavity. In fact, the eastern rim of the crater proved to be beyond the camera’s horizon in the upslope perspective. In an exercise analogous to field surveying, the wealth of detail within the crater enabled a topographic chart of its interior to be compiled – it was an excellent example of how lunar orbiters and landers could work together. The tilt of the lander was measured by a variety of methods and estimated at 12 degrees, inclined almost due west. The fact that this was several degrees steeper than the local slope of 10 degrees was the result of foot pad no. 1, which was on the downslope side, having come to rest in a small depression.

Whereas Surveyor 1’s verniers had cut off as intended at a height of 12 feet and – as hoped – had not disturbed the surface, the fact that Surveyor 3’s verniers had kept firing through two touchdowns offered an opportunity to investigate how an intense gaseous plume affected the lunar surface material. Although the imprints of the first contact were unidentifiable owing to the highly foreshortened view of the rim of the crater and the problem with the camera made it difficult to look for erosional effects beneath the lander, the site of the second contact was conveniently positioned about

A close up of Lunar Orbiter 3 frame H-154 showing the crater in which Surveyor 3 landed, with the inferred position of the lander indicated by the arrow. (The lander was not present at the time, however.)

A contour map of the crater in which Surveyor 3 landed. The contours were drawn by interpolating between control points derived by the photographic trigonometry method. The probable vertical accuracy is +0.5 metre.

“less than 0.5 mm” which was made by tilting the camera’s mirror to view almost directly downward and taking high-resolution pictures of the surface beneath the camera.

After the success of Surveyor 1, it was decided to introduce the soil mechanics surface sampler that had been intended for later Surveyors, possibly for carriage by a rover. The original design included sensors for direct measurement of position, force and acceleration, but because the telemetry and commanding capability of the initial form of the lander could not support this complexity the position measuring system, strain gauges and accelerometers were deleted. Instead, the actions of the arm would have to be monitored by observing it using the camera, and a limited amount of data would be able to be inferred from measuring the current that the motors drew whilst in operation. The experiment comprised the articulation mechanism, the electronics compartment and their supporting structures and electrical cabling. The mechanism was mounted on the space frame immediately to the left of leg no. 2, in the position formerly occupied by the approach TV camera. The electronics compartment was at the same level and almost midway between legs no. 2 and 3. The electromechanical mechanism comprised an arm and a scoop. The arm consisted of tubular aluminium cross members which could be extended and retracted in a pantograph fashion. The scoop was a container about 13 cm long and 5 cm wide, was rigidly affixed to the arm and its door was opened and closed using an electric motor. Three electrical motors operated through drive trains to extend and retract the arm and to rotate it independently in azimuth and elevation. It was spring-loaded, extended by having an electric motor unreel a metal tape, and retracted by reeling in the tape. The arm had a maximum extension of about 1.5 metres, but was unable to access the ground immediately below its mount. It could be elevated to raise the scoop about 1 metre off the ground. On this mission the azimuth arc of the mechanism subtended 112 degrees, ranging from the left edge of pad no. 2 towards pad no. 3 – although because the legs were spaced at 120-degree intervals it stopped short of pad no. 3. In all, the sampler had an arcuate operating area of 2.2 square metres. It was controlled from Earth, but because there was no onboard memory for complex sequences it had to be operated one command at a time. The fact that it used the same radio channel as the camera meant their use had to be interleaved. The principal investigator was Ronald F. Scott, an engineer at Caltech who had worked on the Surveyor rover proposal, but the hardware had been designed and built by Hughes. Of the overall mass of 15 pounds, the mechanical assembly accounted for 8.4 pounds.

The soil mechanics surface sampler, which was also referred to as the ‘scratcher’, could manipulate the lunar surface material in a number of ways. A trench could be made by opening the door of the scoop to expose its blade, driving the scoop into the ground and retracting the arm. The scoop could hold up to 100 cubic centimetres of loose granular material, or a small rock fragment. The mechanism was designed to dig to a depth of 0.45 metre, providing that the material permitted this. An impact test involved raising the scoop and disengaging the elevation motor by releasing the clutch to allow a torsional spring to assist lunar gravity in drawing the scoop down to disturb the surface. If dropped on top of a rock, the scoop’s blade could serve as a rudimentary geological hammer. There was a 2.5 x 5.1-cm strip on the lower edge of

TV TARGET

FOOTPAD г

The configuration of the Surveyor 3 lander.

the door in order to place a flat face on the lunar surface. A static bearing test would involve placing the scoop, door closed, directly above the target and then driving the scoop down until the motor stalled, with the current providing a measure of the force applied. The arm could also be manipulated to push rocks aside in order to inspect either the underside of the rock or its imprint on the surface.

The checkout of the soil mechanics surface sampler started at 10:00 on 21 April, shortly before the end of Goldstone’s second session. After a pyrotechnic was fired to release the mechanism, JPL engineer Floyd I. Roberson commanded the arm to extend. The picture taken to confirm this showed that the arm had not advanced as far as expected. The command sequence was repeated, and the next picture showed that the arm was in the desired position. He then put it through a series of actions to verify that it could move in azimuth and elevation, checking its progress at each step by TV. This done, the arm was drawn back.

On 22 April the arm was swung to the middle of its operating area, and at 05:15 made its first bearing test of the lunar surface. The scoop was raised, the arm was swung to the right, and the scoop, door open, was driven into the surface at 09:14, after which the arm was retracted in order to scrape its first trench. Next the arm was swung left, beyond the bearing test position. After making a shallow scrape, it was raised and repositioned to make a second scrape on the same line. This time the motor stalled after just 10 cm – evidently it was more difficult to scrape an already existing trench. Meanwhile, the camera had suffered a difficulty moving in azimuth

A model of the soil mechanics surface sampler carried by Surveyor 3.

that limited its ability to support the sampling activity. Work on the second trench resumed on 23 April with a third scrape being made along the same line. Arm work was suspended on 24 and 25 April owing to the heat. Because the latitude of the site was 3 degrees, the maximum solar elevation was 87 degrees. The arm had been left at the inner end of the second trench. Having noticed what appeared to be a rock at that position, the team decided to scoop it up on 26 April, but in the process of doing so the object crumbled. The arm was swung as far right as it could traverse and the sample was deposited on the upper surface of foot pad no. 2 so that the camera could inspect the clump of fine-grained material in colour at high resolution.

On 27 April the arm swung slightly left, away from foot pad no. 2, and conducted a second and third bearing test. It then moved a little further left and scraped a third trench involving 26 retraction steps, with a wide-angle picture being taken after each step and later sequenced to produce a ‘stop-motion’ movie. On 28 April the scoop picked up a small bright object from near the most recently made trench. This was added to the material dumped onto foot pad no. 2 for inspection, but there was loose material in the scoop from the trench and on falling from the scoop this covered the white object. When the scoop was dragged across the pile to expose the object of interest, it was observed to have darkened. Next, the arm made two parallel scrapes

The operating area available to Surveyor 3’s soil mechanics surface sampler.

A picture taken by Surveyor 3 on 28 April 1967 showing the soil mechanics surface

sampler positioned between trenches no. 1 and 2.

successively offset to the left of the third trench in order to widen it, and then a bearing test was performed on its floor. On 29 April half a dozen impact tests were conducted in an arc beyond the recent trench, with the scoop being released from a variety of heights in order to vary the force of the impact. The arm was swung to the left of its operating area on 30 April and the scoop manipulated to draw a partially buried bright object onto the surface – it proved to be a fragment of hard rock, and it was photographed in colour. On 1 May two additional scrapes were made to deepen the second trench and then the scoop was dropped four times with its door open to loosen the floor prior to a final scrape. With the Sun sinking in the west the lander’s shadow masked ever more of the arm’s operating area, so on 2 May the arm ended its operations by swinging over to the right to scrape a short fourth trench alongside the broad third trench.

The results of the arm operations indicated that the material was fine-grained and had sufficient cohesion to create loose aggregations up to several centimetres in size, although such ‘clods’ readily fell apart. When the scoop was pressed on the surface for a bearing-strength test, it left a smooth imprint which had a raised ridge of lumpy

V ‘MPACT □ BEARING О CONTACT

material around the edge. This implied that although the material was compressible, it was only moderately so, and after a certain compression the vertical force tended to displace material sideways. Impact tests were performed, but the ‘spring constant’ of the torsional spring proved insufficient to determine the density in this manner. In general, the first scrape of a trench excavated to a depth of about 7.5 cm, and each successive scrape on the same line gained an additional 5 cm – with the arm having to work harder to achieve this. Bearing tests on the floor of a trench showed that the strength of the material increased significantly at a depth of several centimetres. The deepest excavation achieved was about 18 cm, which was less than half of that for which the arm’s range of operation had been designed. Nevertheless, it provided a valuable insight into the third dimension of the enigmatic fragmental debris layer. There was no indication of textural layering in the walls of the trenches. If there was any change in the grain size, this was on a scale finer than the camera’s resolution. It simply seemed that the upper few centimetres were porous, and hence compressible, whereas the essentially similar material below was more consolidated. Its cohesivity was confirmed by the fact that the trench walls did not collapse. As in the case of the material disturbed by the foot pads, the subsurface was significantly darker than the undisturbed surface – in retrospect, it was more as if the uppermost few millimetres had somehow been lightened. From the fact that no bright angular fragments were uncovered in trenching, it was speculated that while buried they became coated with dark fine-grained material and in this darkened condition were difficult to see in a trench. By implication, it seemed that after a rock had been exposed on the surface for a time it was ‘cleaned off’ by some form of weathering. There were only a few rocks within the arm’s operating area. Most were small and partially buried. The arm picked up one rock for a close examination, but it was too small for its mass to be measured. The jaws of the scoop picked up a small white rock which was about 1.2 cm in size – a task involving 90 minutes of careful remote-control manipulation. The 100-psi pressure which the scoop exerted would have crushed a weak terrestrial rock such as a siltstone or friable sandstone, but the lunar rock remained intact.

The lesson for Apollo was that whilst the lunar material was very fine-grained, it was moderately cohesive and its bearing strength increased significantly at shallow depth.

Surveyor 3’s view was confined to the 200-metre-diameter crater in which it had landed – it could not see the plain beyond. The craters in view ranged in size from 10 cm up to 25 metres. Most of the craters that were less than 3 metres in diameter were fairly shallow, and either had very subdued raised rims or were rimless. Most of the craters between 3 metres and 12 metres in diameter were subdued, but 25 per cent had raised rims and relatively steep walls. It was apparent that most of the small craters had not penetrated beneath the fragmental debris layer within the main crater, and had merely redistributed the material that was already exposed at the surface. The size-frequency distribution was similar to that for this size range seen on the maria by Rangers 7 and 8.

The angular-to-rounded fragments ranged in size from tiny grains up to blocks of about 1.5 metres. The albedo of the undisturbed surface was 8.5 (±2) per cent, and in some cases the albedo of the blocks was one-third brighter. Although the camera

operated most effectively when the Sun was high in the sky, in such illumination the absence of shadows made subtle terrain relief almost impossible to discern – but on the other hand in such illumination it was straightforward to chart the distribution of blocks. Most blocks were relatively angular, with many wedge-shaped and some even tabular. Some of the angular rocks were partially buried, but most of the well – rounded fragments were fairly deeply buried.

In addition to the sparse and random litter of blocks, there were two prominent ‘strewn fields’ of coarse blocks. One was clearly associated with a sharp raised-rim crater about 13 metres in diameter that was embedded in the northeastern rim of the main crater, some 80 metres from the lander. The other was associated with a pair of subdued craters that were located high on the southern wall. The line of sight provided a view inside the northeastern crater, revealing its interior to be full of similar blocks. Exterior to the rim, there were radial lines of blocks. The blocks associated with this crater were the largest, coarsest and most angular in the lander’s field of view. It was evident that they were ejected by the impact that created the crater, and derived from material at a depth of 2 or 3 metres beneath the rim of the main crater. Some of the blocks had almost planar faces, as though they had broken along pre-existing joints. The tabular ones displayed grooves and ridges on their narrowest sides suggestive of lamination parallel to their longest dimension, such as would be produced in flow-banded lavas. The blocks associated with the southern craters were of similar size, but were more rounded and tended to be more deeply buried. Their source was probably the larger of the two craters there, which was 15 metres in diameter. These observations suggested that large blocks associated with subdued craters tended to be more rounded than those associated with sharp raised-rim craters, and those around subdued craters were more buried than those of sharp raised-rim craters. This suggested that freshly exposed blocks were not only eroded by the rain of meteoritic material, but also tended to be reburied as material accumulated – either by the arrival of further ejecta or as a result of downslope motion of loose debris. The fact that the rounded blocks had a pitted texture whereas the angular ones did not, implied that the pitting was caused by the same process that rounded off the angular blocks.

It was also apparent that the surface on the interior of a sizeable shallow crater on a mare plain was similar to that on a relatively level area between such craters. The 200-metre crater in which Surveyor 3 landed had probably been partially filled in by the downslope motion of material on its interior walls, thickening the debris towards the centre. Loose material piled up against the upslope sides of the larger blocks was interpreted as evidence of this process. The strewn fields of coarse blocks associated with 13-15-metre-diameter craters on or near the rim of the main crater implied that the fragmental debris layer was about 2 metres thick there. In contrast, the fact that a 20-metre crater near the centre of the main crater had not excavated blocks served to confirm that the layer there had been thickened. When the main crater was created, its floor would have been several tens of metres deeper and its rim several metres higher and sharper than it is today. By the ‘hinge-flap’ effect of an impact, the debris that formed the rim would have been excavated from the deepest point. In effect, a blocky crater on the rim of a larger crater serves as a ‘drill hole’. The crater on the

A northward-looking section of a panorama taken by Surveyor 3. The outline shows the area covered by the next illustration. (Courtesy of Philip J. Stooke, adapted from International Atlas of Lunar Exploration, 2007)

A portion of the previous illustration featuring a strewn field of boulders around a small crater on the northeastern rim of the crater in which Surveyor 3 landed. (Courtesy of Philip J. Stooke, adapted from International Atlas of Lunar Exploration, 2007)

northeastern rim would offer visiting astronauts an opportunity to recover material excavated from beneath the fragmental debris layer.2

After a detailed analysis of the Surveyor 3 imagery, Gene Shoemaker introduced the term ‘regolith’ to lunar science. This was familiar to terrestrial geologists as the collective name for the rock wastes of whatever origin and however transported that rest on bedrock and nearly everywhere form the surface of the land. On Earth, there are many erosional processes and the regolith includes volcanic ash, glacial drift, alluvial deposits, eolian deposits and soils rich in humus. On the Moon, the primary erosional process was meteoritic impact. Harold Urey had introduced ‘gardening’ for the manner in which the poorly sorted fragmental debris layer was turned over by impacts. The pictures from the two Oceanus Procellarum landing sites hinted that the thickness of the layer increased with age. To start with, the surface would have been bare rock. Any significant impact would have been capable of breaking up and scattering the material. This ejecta would have been progressively eroded by the rain of smaller projectiles. Over time, the layer of debris would have thickened, requiring ever larger impacts to reach the substrate. An important aspect of this process was that it would yield a continuous distribution of fragment sizes, which was expressed by saying that the lunar regolith was seriate.

One scientific task for Surveyor 3 was to monitor the Moon’s passage through the Earth’s shadow. With the Moon at ‘full’ phase and the Earth masking the Sun, this was a lunar eclipse to a terrestrial observer and a solar eclipse to the lander. It was the first opportunity to observe the thermal effects of such an event from the lunar surface and assess inferences drawn from telescopic studies. In particular, Surveyor science team member John M. Saari had been involved in infrared scanning of the Moon’s disk during the lunar eclipse of 19 December 1964, the data from which was processed into isothermal contours that indicated the presence of many ‘anomalies’, mostly associated with craters, where the heat that had been absorbed while the Sun was shining at lunar noon was radiated again at the onset of the eclipse. In addition to monitoring the temperature, Surveyor 3 gave the scientists a bonus: the mirror of its TV camera had the same 35-degree elevation limit as previously, but because the lander was west of the meridian and inclined due west at an angle of 12 degrees by virtue of having settled on a slope, the field of view of the wide-angle frame was just able to include Earth, east of the zenith. Optical observations of the eclipse would yield the first direct measurement of the distribution of the refracted sunlight which weakly illuminates the lunar disk at such times.

Fortunately, the eclipse on 24 April occurred during a period when the Moon was still just above Goldstone’s horizon. A total of 20 images were taken in two sets: the first at 11:24 and the second 37 minutes later. They were taken at two iris apertures, and with several exposures for each of the three colour filters. In the first set, an arc along the northwestern limb of Earth refracted light that varied greatly in brightness, and with a fainter glow at each end containing bright ‘beads’. In the second set, the

And in fact Apollo 12 would do so at precisely this spot.

• INFRARED ANOMALIES

The locations of lunar transient events reported over the years by various observers and infrared ‘hot spot’ anomalies measured during a lunar eclipse on 19 December 1964. (Courtesy of John M. Saari and R. W. Shortfall, Isothermal and Isophotic Atlas of the Moon, NASA, 1967)

brightest refraction had migrated to the northeastern limb. To assess the distribution of cloud on the limb, the geographic coordinates of the ‘beads’ were later calculated and compared to pictures taken by the ESSA 3 meteorological satellite in low polar orbit on the day prior to the eclipse. In some areas cloud in the troposphere occulted some of the refracted sunlight, but the bright ‘beads’ occurred where sunlight passed through regions free of cloud. Of course, the refracted sunlight made it impossible to view the much fainter solar corona.[34]

The thermal properties of the lunar surface inferred from the lander’s temperature data differed from inferences made from telescopic studies, in that the in-situ data showed a higher thermal inertia. However, the data was provided by sensors in place to monitor the thermally controlled compartments, not by instruments specifically designed to study the thermal properties of the lunar surface, and therefore was too crude to draw definitive conclusions.

The plan had called for attempting a ‘liftoff and translation’ manoeuvre by firing the verniers, but this was ruled out by thermal factors. The decision rested upon the temperatures of the thrust chambers of the engines, the flight control electronics, the helium tank, the shock absorbers and the roll control actuator on vernier no. 1 – all of which depended upon solar heating and shadowing in the orientation in which the vehicle came to rest. The key issue was the temperatures of the engines. Irrespective of the orientation, by the time the elevation of the Sun reached about 35 degrees one or other of the engines was sure to exceed its permitted pre-ignition temperature of 105°C. Hence, any attempt to lift off had to be made either early in the morning or late in the afternoon. For Surveyor 3, the thermal situation was complicated by the fact that the lander was on a slope, which significantly altered the manner in which shadows were cast. To preserve the option of this experiment, the helium tank had not been vented. Also, given that at no time during the protracted landing had the forces on the shock absorbers imparted even half the load endured by Surveyor 1, it had been decided not to lock the legs. Whilst the helium was significantly depleted, as the tank absorbed solar heat its pressure increased to 2,735 psia.[35] By the time of the eclipse, however, it had been decided not to attempt to fire the verniers. The legs were monitored to determine whether the rapid decline in temperature at the onset of the eclipse prompted the shock absorbers to leak, but they retained their integrity. At 20:36 on 24 April, the helium tank was finally vented.

Surveyor 3 had two flat beryllium mirrors situated to enhance the camera’s view of the underside of the vehicle. That is, the camera was between legs no. 2 and 3 and the mirrors were affixed to leg no. 1. One mirror was 35 x 22 cm and gave a view of the lower portion of crushable block no. 3 and the area beneath vernier no. 3. The other mirror was 9 x 33 cm and viewed the area beneath vernier no. 2. The fact that the verniers were cut off at a height of only 3 feet suggested there might be signs of surface erosion, but since the vertical velocity at the third contact was only 1.5 ft/sec the crushable blocks probably did not strike the surface. Unfortunately, the hopes of viewing beneath the lander were foiled by the fact that when the Sun was low in the sky the pictures were ‘washed out’ by the glare from the coating on the main mirror of the camera, and when the Sun was high in the sky the area of interest was in the lander’s shadow! It had been hoped to fire the downward-pointing cold-gas thruster on leg no. 2 to follow up on Surveyor 1’s surface erosion experiment, but this time obtaining good ‘before’ and ‘after’ pictures of the test area. However, pictures taken when the Sun was low in the east were washed out by glare, pictures taken when the Sun was high lacked the shadows required to highlight the minute changes likely to result from such a weak thrust, and in the late afternoon the area was in the lander’s shadow – so the test had to be cancelled.

Between 10:29 and 11:06 on 30 April, Surveyor 3 snapped wide-angle pictures of Earth illuminated as a crescent with the dawn terminator running the length of South America. These were the first colour pictures of Earth taken from deep space.[36] The filters had been revised to provide an improved spectral match to standard colorimetry functions.

Following sunset at 18:38 on 3 May Surveyor 3 monitored the rate at which the temperature fell, and at 00:02 on 4 May was commanded to hibernate.

In total, the camera took 6,326 pictures. Owing to the glare from contamination of the main mirror, usable images could be obtained only over a limited azimuth range during the early morning and late afternoon. This glare, combined with the difficulties in moving the mirror, made it impossible to obtain all of the systematic surveys of the landscape which had been planned. About 8 per cent of the pictures were taken at wide-angle to provide panoramas at specific illumination phases. The glare impaired detailed photometry, but the colorimetry confirmed that the surface was essentially grey. The glare precluded photographing the stars for use as celestial references to precisely determine the orientation of the camera, and hence the true orientation of the lander, but on one occasion it did manage to photograph Venus, which helped to some extent.

Although Surveyor 3 was unable to view the plain surrounding the crater in which it landed, scientists were delighted to have the opportunity to survey the interior of a medium-sized crater on a mare! In effect, therefore, its observations complemented those by Surveyor 1 of the open plain. In general, the character of the lunar surface material appeared to be similar at the two sites. The soil mechanics surface sampler was active for a total of 18.3 hours. It executed 5,879 commands, during which it made seven bearing-strength tests, thirteen impact tests and four trenches to provide data on the strength, texture and structure of the lunar material to a depth of 18 cm. In particular, it found that the bearing strength of the material increased with depth, even although there was no discernible change in the grain size – it was just that the uppermost few centimetres were more porous. Although sequences of commands to enable the sampler to perform complex operations had been stored on magnetic tape for step-by-step uplinking, for much of the time it was actually operated in real-time and monitored by TV.

Surveyor 3 evidently succumbed to the chill of the lunar night, because attempts to reactivate it after sunrise were unsuccessful.