The Design and Engineering of Curiosity

by Emily Lakdawalla

  5.4.5 Cached sample operations and doggie bagging

  At Rocknest it quickly became apparent that if the rover couldn’t drive while CHIMRA held sample, the mission could be stuck with a prolonged delay. The SAM team wanted to take many deliveries of sample, running their experiment in different ways each time. One option would be to perform many dropoffs to SAM sample tubes before driving away, a procedure called “doggie bagging.” Curiosity does doggie-bag samples, but not many; the SAM team has to strike a balance between holding options open for future analyses and consumption of clean sample tubes.

  The engineers came up with a workable solution: driving with cached post-sieve sample in the stowed turret. The clamshell happens to be positioned perfectly for long-term sample storage (with its opening upward) when the turret is stowed. There is sufficient room in the clamshell for 12 cubic centimeters of sample without spilling, provided that the rover’s tilt does not exceed 20°. The limitations exist because sieved sample must not be allowed to fall onto the back side of the 150-micrometer sieve. Thwacking is a one-way operation designed to motivate particles out of the sieve toward the reservoir. Any particle that falls on the back of the sieve and clogs a hole would be further embedded by thwacking and likely stuck forever. After performing any cached-sample contact science with the arm, CHIMRA does a recovery sequence of rocking and vibrating to ensure that any sample that may have escaped the clamshell returns to it before the arm is stowed.

  The limits for cached sample operations were developed very quickly, early in the mission, with many time pressures on the engineering team. Cached-sample operations required lots of extra arm motions to move the sample to different locations depending on the desired orientation of the turret, always preventing the sample from falling on the back side of the sieve. Later, the engineering team developed a new set of operational rules called evolved cached sample operations, which they first used on Mars on sol 1064. While protecting the safety of the CHIMRA system, the new rules allow some sample to fall on the back side of the sieve, requiring fewer arm motions and therefore less energy and time to run. This increases the flexibility of cached-sample operations.15

  5.4.6 CHIMRA concerns and anomalies

  CHIMRA has mostly functioned as designed – quite a coup for such a complicated piece of equipment, the likes of which has never before been sent to another planet. There are two issues affecting its use: concern about its 150-micrometer sieve, and a problem with the primary thwack actuator.

  5.4.6.1 150-micrometer sieve edge weld separation

  Originally, three identical CHIMRA devices were built. One is now on Mars, one is on the testbed rover, and one was tested to failure on Earth in the Qualification Model Dirty Testbed (see section 4.​7.​4). After performing about 130 primary thwacks with the test unit, the edge welds that hold the 150-micrometer sieve onto the primary thwack arm began to pop apart, creating a gap through which larger particles could leak into the sieved sample.16

  To prevent the degradation of the edge welds on the flight unit of the 150-micron sieve, engineers now command primary thwacks only when preparing for a new sample. After some primary thwacks, they use ChemCam to perform a detailed inspection of the edge welds and the sieve, including angling the sieve to direct a specular reflection at the cameras in order to search for any deformation (Figure 5.14). Table 5.1 lists all sols on which ChemCam inspection of the sieve was performed.

  Figure 5.14. ChemCam RMI inspection of the 150-micrometer sieve performed on sol 1048. NASA/JPL-Caltech/CNES/CNRS/LANL/IRAP/IAS/LPGN/mosaic by William Rapin.

  5.4.6.2 Drill sample cross-contamination

  Transfer of material from drill to CHIMRA was originally intended to be done with some drill percussion. Following the sol 911 drill percussion anomaly (section 5.3.4.2), engineers developed a method to perform sample transfer with limited percussion. However, continued testing suggested that this new method did not empty the drill sample chamber as effectively as before, increasing the risk that sample material from a previous site might cross-contaminate a new sample. Indeed, CheMin results suggest Buckskin sample cross-contaminated the Big Sky sample, although there could be other explanations (section 9.​4.​4). Engineers experimented with a new non-percussion sample transfer method on Mars on sols 1460 and 1494 to reduce cross-contamination risk.17

  5.4.6.3 Primary thwack actuator anomaly

  On sol 1231, during routine sample processing, the primary thwack arm stalled. After sieving, Curiosity typically cracks open the primary thwack arm to peek into the portion box to estimate how much sample is inside. This time, it stalled after opening only 1°. Cautious testing since then has seen the actuator operate fairly normally without stalling. (They tested cautiously because if the primary thwack arm were to fail in a wide-open position, it would no longer be possible to sieve a fine fraction, making it difficult to prepare samples for delivery to CheMin.) Testing included extra imaging of primary thwack arm motion on sols 1237–1243.

  Engineers have reduced use of the primary thwack arm in order to avoid faults. They no longer crack open the primary thwack arm to inspect the post-sieved sample. Since Lubango (drilled sol 1324), they drill, inspect the pre-sieve (coarse) fraction of the sample in the scoop, immediately dump this fraction, transfer the post-sieve (fine) fraction to the scoop, and visually inspect it there, thereby shifting the workload to the secondary thwack actuator rather than the primary thwack actuator.

  5.5 DRT: DUST REMOVAL TOOL

  The dust removal tool is a brush for cleaning rock surfaces before they are studied with MAHLI, APXS, or Mastcam’s narrowband filters. It consists of two wire brushes, driven by a single motor (Figure 5.15).18 When Curiosity landed, functions related to the use of the brush had not yet been tested on Earth, so its first use was on sol 150. Initially, rover planners inspected the brush only after brush operations. On sol 291, following the third brushing operation, at Cumberland, they discovered that one set of bristles was bent (Figure 5.15, middle row), leading to concern that the bent wire bundle could wrap around the brush’s central spindle during brush operations. The mission halted use of the brush and began a period of extensive Earth testing. The brush was not cleared for use on Mars until arrival at the next drill site, Windjana, and didn’t see routine use until arrival at Pahrump Hills. Since then, they have imaged the brush both before and after each operation, and no further degradation in the brush bristles has been observed (Figure 5.15, bottom row). Table 5.2 lists all brush sites up to sol 1514.

  Figure 5.15. Condition of the dust removal tool (DRT) over time, as seen in standard right Mastcam engineering support imagery. Top row: after its first use on sol 150. Middle row: after the third use, at Cumberland, when one wire bristle was discovered to be bent. Bottom row: after more than 50 uses, on sol 1416, at Chibia. NASA/JPL-Caltech/MSSS.

  Table 5.2. Summary of dust removal tool (brush) to sol 1806. List courtesy Ken Edgett.

  Bradbury Group

  Pahrump Hills

  North of the dunes

  Among the dunes

  South of the dunes

  150 Ekwir

  169 Wernecke

  291 Cumberland

  612 Windjana

  722 Bonanza King

  755 Maturango

  758 Moenkopi

  767 Morrison

  805 Pelona

  806 Ricardo

  808 Rosamond

  809 Mojave

  813 Punchbowl

  814 Afton Canyon

  815 Topanga

  819 Mescal

  824 Puente

  828 Pickhandle

  830 Goldstone

  844 Santa Ana

  853 Tecoya

  880 Mojave 2

  905 Telegraph Peak

  936 Hyrum

  975 Albert

  998 Ronan

  999 Wallace

  1057 Buckskin

  1092 Ledger

  1105 Winnipeg

  1109 Cody

  1114 Big Sky


  1119 Big Sky 2

  1130 Greenhorn

  1157 Augusta

  1166 Swartkloofberg

  1245 Kudis

  1251 Kuiseb

  1259 Gorob

  1266 Stockdale

  1273 Schwarzrand

  1275 Mirabib

  1279 Khomas

  1287 Sesriem Canyon

  1293 Brukkaros

  1300 Bero

  1318 Lubango

  1330 Okoruso

  1341 Kwakwas

  1348 Meob

  1355 Inamagando

  1358 Oudam

  1366 Auberes

  1380 Koes

  1416 Chibia

  1418 Marimba

  1436 Conda

  1444 Ganda

  1474 Jwaneng

  1477 Catumbela

  1484 Serowe

  1491 Sebina

  1511 Penobscot

  1523 Sutton Island

  1531 Precipice

  1586 Belle Lake

  1681 Duck Brook Bridge

  1695 Mason Point

  1702 Fern Spring

  1710 White Ledge

  1715 Timber Point

  1736 Winter Harbor

  1744 Mingo

  1806 Robinson Rock

  The motor speed can be changed during a single brush operation, and the robotic arm can be moved while the brush is running to sweep an elongated area. If Curiosity were to hold the brush in one position during use, the brushing action would leave an unbrushed spot in the center. So Curiosity moves the arm while brushing to sweep out the center, creating an oval spot. The entire brushed area is contained within a 60-millimeter circle.

  Figure 5.16 shows two different kinds of brushed spots. The size of the brushed spot depends upon how close the brush gets to the surface. Originally, Curiosity brushed at a height of 1 centimeter, which produced a brushed spot about 46 millimeters in diameter. Concern that the bristles could wrap around the center post following the discovery of the bent bristles led them to use a higher standoff of 1.5 centimeters thereafter, which produces a brushed spot about 40 millimeters in diameter. Either way, the cleared area is slightly wider than the field of view of APXS, so even with positioning uncertainty, APXS’s field of view will be entirely in the brushed area.

  Figure 5.16. Moenkopi (left), brushed on sol 758, was a raised feature. Maturango (right), brushed sol 755, was a flat spot, and the brush was moved during brushing. The Confidence Hills drill site is in the background. Drill holes are 16 mm across; brushed spots are at least 45 mm across. Left Mastcam image 0758MH0001900010204611C00. NASA/JPL-Caltech/MSSS.

  When the brush interacts with very soft rocks, the wire bristles may scratch the surface. If the rock is extremely soft, wires near the center can get hung up on a protuberance and actually drill into the rock (see Figure 5.17 for an example).

  Figure 5.17. Scratches and drilling near the center of a brushed spot at Pelona, at the Pahrump Hills site, sol 805. MAHLI images 0805MH0001900010300492C00 and 0805MH0003070010300514C00. NASA/JPL-Caltech/MSSS.

  5.6 ORGANIC CHECK MATERIAL

  A palette on the front center of the rover contains 5 cylinder-shaped bricks of hermetically sealed organic check material provided by the SAM team (Figure 5.18).19 It is intended to check the cleanliness of the whole SAM sample processing pathway, to ensure that any organic materials detected by SAM in Martian material really do come from Mars and are not Earthly contamination left behind on the external parts of the sample processing chain: the drill, CHIMRA, and sample inlets. The five bricks are identical. They are composed of a ceramic that has been doped with a minute amount of fluorinated hydrocarbon chemical that can be detected by SAM. Each brick is covered with a foil seal. It can be drilled and sampled just like a rock, and the sample dropped into SAM. Drilling it breaks the seal, so each brick can be used only once. Figure 5.19 shows how the rover would position the drill on one of the bricks for sampling.

  Figure 5.18. The organic check material mounting plate bolted to the front center of the rover contains five foil-capped cylinder-shaped bricks of ceramic material. Between the five bricks are smaller drill positioning pads, places for the drill contact stabilizers to press during drilling. Below the mounting plate is one of the two spare drill bit boxes and the four front Hazcams. Mosaic of four MAHLI self-portrait frames taken on sol 1065. NASA/JPL-Caltech/MSSS.

  Figure 5.19. A test of the drill’s positioning on one of the organic check material bricks. From this point of view, you can see the tubular inlet and outlet ports that allowed the SAM team to dope the ceramic bricks with a fluorinated compound after they were sealed in their can-shaped housings. Navcam image NLB_493020688RAD LF0490814NCAM00467M1, sol 1076. NASA/JPL-Caltech.

  The organic check material has not yet been used. (For more information, see 9.​5.​1.​12.) The team tested the process of positioning the drill to sample the organic check material on sols 34 and 1076, taking images with MAHLI to document turret positioning (Figure 5.19). They also performed imaging of the organic check material with the MAHLI cover closed on sol 1416.

  5.7 SAMPLE PLAYGROUND

  Immediately in front of the mast is a suite of tools intended to allow the rover to study the properties of drilled or scooped sample. This “sample playground” was a late addition to the rover design, added after the experience of the Phoenix mission, which had a difficult time handling clumpy Martian materials (see section 1.​5.​11).20 The sample science team looked at how their mature design might be vulnerable to clumpy soil, and developed a set of tools to help them investigate sampled material before committing to delivering it to CheMin and SAM.

  The sample playground comprises the science observation tray, engineering tray, CheMin surrogate funnel and soil capture plate, dust removal tool scratching post, and two portion pokers (Figure 5.20). Sols in which science cameras were used to image parts of the sample playground are listed in Table 5.3, but the playground is also often captured in Mastcam and Navcam views of the work volume, and in MAHLI self-portraits.

  Figure 5.20. Parts of the sample playground seen from above by Mastcam and from the right side by MAHLI. In the top image, two portions have been dropped to the center of the observation tray. CHIMRA vibration transferred through the arm to the rover body has caused the first portion to “walk” off the tray during delivery of the second portion. Delivery of both portions has also shifted some of the accumulated dust off of the tray. Mastcam image ML0005780000102730E01 and MAHLI image 0544MH0003460000201460C00. NASA/JPL-Caltech/MSSS/Emily Lakdawalla.

  Table 5.3. Sols with targeted imaging of the sample playground (mostly of the observation tray).

  Mastcam

  MAHLI

  70

  76

  77

  78

  79

  81

  95

  284

  289

  37

  73

  93

  95

  177

  544

  571

  572

  The science observation tray, also known as the “O-tray”, is a circular titanium plate 75 millimeters wide. The rover can drop portions of drilled or scooped sample onto it for investigation with APXS, MAHLI, and mast-mounted cameras. It was intended to provide a surface of known composition on which to perform APXS observations of sampled material. Unfortunately, vibration from CHIMRA, necessary for portion delivery, appears to transfer through the rover arm to the body and cause delivered portions to “walk” off of the sample tray (see Figure 5.20 for an example). This behavior, which is much worse on Mars than it was in the Earth testbed, has prevented much use of the observation tray for science. The APXS team has developed a different method of measuring the composition of drilled materials by studying the composition of pre- and post-sieve dump piles. The APXS team has also occasionally taken advantage of the vibration-induced cleaning of the observation tray to perform measurements of the composition of airfall dust.

  The rest of the sample pla
yground elements have not been used on Mars. The engineering tray, located closer to the rover body than the observation tray, has a checkerboard pattern made of 0.25-inch (6.35-millimeter) squares. It was intended for use in estimating the volume of portions dropped to SAM and CheMin. To the left is the CheMin surrogate funnel and soil capture plate. If there are concerns about soil clumping and potentially clogging the CheMin inlet funnel, the drop can be tested with the surrogate funnel. On the right side of the engineering tray, a palette of screw heads provides a place to clean off the DRT in the event that Martian material clings to its brushes. Finally, two portion pokers, one pointing vertically and one horizontally, provide Curiosity with the capability to poke out the CHIMRA portion hole if it should become clogged with material. However, the inverted funnel shape of the CHIMRA portion hole makes it very unlikely that material could pass all the way through CHIMRA and then clog the portion hole; the portion pokers have not been used and hopefully never will be.