Curiosity

by Rod Pyle

  Now, remember that this drill is mounted on the turret on the end of the arm along with the MAHLI camera, the APXS instrument, and other devices, so it can't be allowed to beat things up too badly on the turret while it's running. It was insulated enough to avoid damaging anything nearby. And on top of all of this, it had to be ultrareliable (like everything else) and lightweight enough to fly into space.

  The engineers designed, built and tested, and tested some more. As they learned of the limitations of the design and its weaknesses, they reengineered it, rebuilt it, and tested it again.

  Late in the game, shortly before launch, in fact, they discovered that there was a problem. The drill could cause a short circuit, which would theoretically take down a good part of the rover's electronics. Nobody I spoke with called it a panic, but it had to come close. Soon after the engineering team had a fix, Rob Manning was working on it, jiggering it in place, while the rocket was actually on the pad waiting to launch. It was that close.

  I'm going to let one of the people responsible for this fix action narrate the process. We met Dan Limonadi during our tactical-planning meeting up at JPL; he was the Captain America–type I droned on about: tall, lean, just a hair over forty, and a community icon outside of work. A large part of his responsibility on MSL was the arm and its instrumentation, which includes the drill.

  Fig. 27.1. ALL BUSINESS: This image of Curiosity's robotic arm shows the drill (as indicated by the arrow), facing the camera in mid-turret. The bit is a jagged, percussive design. This photograph was taken in Yellowknife Bay. Image from NASA/JPL-Caltech/MSSS.

  “In the late summer of 2011 we were doing a lot of testing with Curiosity's sampling system, and we had multiple issues that hit us all at once. One of them was a short in the arm and the drill, so we scrambled to see what was going on and it was a very intense time.” See? No panic, just intense. He did not mention that a lot of twenty-hour days were involved, but they were. “We came up with a quick fix that would allow us to operate in the face of an electrical short should it occur. We had to come up with a design, review it with a bunch of people, and then work it in at the Kennedy Space Center with an already-buttoned-up spacecraft.” That is as last-minute as it gets. “That was a very exciting time in [a] not-such-a-good way. We actually discovered the fault on the engineering model [the nonflying twin of Curiosity]. This was a funny twist of fate.” Again, with the understatement.

  “In the drill assembly there is a spring attached to the voice coil and the percussion drill. This hits the percussive mechanism. The coil uses magnetic force to move the hammer back and forth, so the spring gives it a bit of a natural frequency. It turns out that the spring in the engineering model was a little bit out of true. The team knew that when they put [it] together, they knew they didn't quite have the right spring that they needed to get testing going, but we were short on time. It turns out as we when were test drilling with that spring, the side started rubbing and it caused a failure—one that we would never see in flight with the proper spring. It didn't actually short the wires inside the drill, but it did expose some of the wires.” So the failure did not actually occur, but in theory it could—just not on the final, flight version. But you simply cannot be too careful. “Once again, this is something that would never happen in flight with the better parts.” All good and true, but it's not enough to let the engineers sleep at night. It had to be fixed.

  Fig. 27.2. THE DRILL EXPLAINED: This schematic view of Curiosity's drill shows (from bottom left) the sheathed drill bit—the sheathing encourages powder from the rock being drilled to crawl up the bit traces into the collection area; the first chamber where rock powder collects after crawling up the bit; the second chamber where powder migrates after being stored in the first chamber; and finally (on the right) the exit orifice to send powder to the analysis instrumentation. Image from NASA/JPL-Caltech.

  Limonadi continues: “But what it did do was expose the class of fault just in time to do something about it before takeoff. It was a minor miracle and a bit of fate, good and bad timing. It was funny in retrospect that [it] happened when it did. So we might [have] been in a different situation if we had [to] operate that drill without having had a chance to put in some mitigation.”

  “As we dug into this, we started testing other units more rigorously, and they failed in a different way, but with the same result of shorting out the chassis. So we learned a lot of different things that we had fix, to do proper mitigation. We discovered that we could actually induce a short in the arm, sending current into the wrong place by choice, and detect every bit of rattling so that if need be we could turn off the drill before it gets damaged.”

  In short (no pun intended, I swear), they did a couple of things to fix the problem. First, they made very sure that the spring on the flight unit was perfect and would not cause the first problem they had detected. Then Manning and his team designed devices to better secure anything else that could cause a short in the drill assembly. They also figured out how to ensure that if they did see any issues during operations on Mars, they would be able to shut the drill off before it could hurt anything else. There was also some discussion of how to route the electrical current to avoid any damage in the unlikely event of a short, but that is above my pay grade and likely above your level of interest. Suffice it to say it was fixed.

  It is worth mentioning that the drill tests that exposed the short were far beyond the expected use of the drill over a period of years. We are talking literally tens of hours across many weeks of use at the top setting. This is way more drill time than the entire primary mission of one Martian year (two Earth years) will utilize. But that is how these guys roll—test for maximum effect, simulate the worst-case scenario, and fix everything possible. Now you know why Opportunity has been roving Mars for over a decade—this level of dedication pays off.

  The expectations for drilling activities were also downgraded about this time, so the demands were lessened. This changed the target run time for the drill. Ashwin Vasavada remembers when NASA changed the number of expected drill uses for the mission. It had been a number that, as drill testing proceeded, was starting to make a lot of people uncomfortable. It was up in the seventy-hour-plus range. Then—“NASA changed that requirement for us to twenty-eight [hours]. When we actually mapped out how many calendar days there are in the two-year mission, how long does it take to get through all the events we had to accomplish, twenty-eight was the right number in terms of the acquisition for the drill. So that's what the system had to be designed to do.” Whew—a break at last. The drill would only be expected to function for about a third of the original time specified. But even this new number—twenty-eight hours—was a lot for a mechanical device as violent as a percussive drill. “You're talking about twenty-eight times you're going to put a drill into a rock. The rover could slip and the drill bit could get stuck, all those concerns. At that point, you don't know what the rock will be made out of, it could turn into goo and clog everything up.” And that's just grabbing the sample. Then you need to investigate it. “Then you need to run those twenty-eight samples, and that means putting them in a mass spectrometer with [one] part per billion accuracy! Talk about not having cross-contaminations between samples. For a long time, I don't think that anybody thought it would be possible. It's not even just drilling robotically—with a science sample, even if you did that in a university lab setting, people would be wearing a gown and gloves and all that. So that's what we had to invent.” It was their own little biohazard-level-4-caliber roving clean lab.

  And there was still another bug in the ointment: Teflon.

  Vasavada explains: “We also had a Teflon contamination issue. This came up in the same timeframe as everything else, between September and November of 2011. This is a really challenging issue because you have a complex system with a lot of moving parts designed to break the rock you are trying to sample. It's also trying to hold onto and collect the powder the drill creates. So there is a can, two
chambers, in the front of the drill, right next to the bit, to collect the powder. To do this, there's an exit perpendicular to the drill.”

  It helps to understand that a tiny bit of rock powder was supposed to crawl up the spiral on the drill bit and land in the collection chambers—the CHIMRA unit—but it was impossible to know what angle the entire assembly would be in when it was drilling. It might be drilling straight down, or to the side, or any angle in between. “We were supposed to be able drill anywhere from vertical to horizontal, and remember that the robotic arm has five degrees of freedom—a human arm has seven degrees of freedom by comparison. So if a science team member wants to place the drill against a horizontal target, we can't control the orientation of the little exit tube.” The drill could, in theory, be in almost any orientation when in operation, except possibly upside down. And no matter what angle it was operating from, the powder it generated needed to travel up the bit, collect in a chamber, and stay there long enough to be processed.

  The answer was to have two collection chambers so that the second chamber would collect the powder from the first before it spilled back out. It would take a bit of clever arm choreography and a lot of testing to make sure that it worked in all possible orientations, but they figured it out.

  But what about the Teflon? Dan Limonadi takes the story back up: “Recall that this is a percussive drill and you need get the hammer force to the bit. Around that bit at the top, these chambers need to have a flexible interface which is thin titanium with a small circular clamp holding it; it's a friction fit around the edges. That seal was Teflon-impregnated fiberglass. So the issue we had was, you hit the diaphragm thirty times a second and there is no way to avoid chafing some of that Teflon-impregnated fiberglass off. And the SAM instrument is supersensitive, and you have to make sure that you don't get a mote of dust or something from the sampling system in there that is not supposed to be there. It's almost an impossible job. The tolerances are incredibly small.”

  SAM measures things at the molecular level, and anything that was not Mars dirt was a no-no and could screw up the readings. The engineers had another talk with the science team as they were trying to design something that seemed to be impossible.

  “In the final analysis, it ended up not being that big a deal once we were on Mars. The SAM team concluded that the small amounts of Teflon they might see could be a little annoying, but it's not the end of the world. [They] didn't expect to see much in the sample. So for launch were thinking about the worst-case scenario and wondering if that was going to be a problem. That caused a big stir, but ultimately the SAM people said it was going to be fine.”

  Also, for the most part, Teflon in the sample could be accounted for when the results came down from SAM anyway. “Teflon is a chlorofluorocarbon, and there are no natural compounds that have that chemistry, so there wasn't really a risk of a false positive. You can tell this is a man-made carbon compound. There was, however, a concern that if there was enough Teflon mixed in, it could possibly mask some of the natural Mars-originated materials found in the samples. If you have ten parts per million of Teflon and one part per million of Mars organics, the Mars organics might be hidden by the Teflon. But in the two drill samples we've done so far, the SAM team says they have not found any Teflon in the sample.”

  He summed it up in a way that could describe so much in the MSL mission, where any errors or faults could be amplified by an exponent over anything that had gone before. “There was an interesting [there's that word again!] pucker factor, and everybody needs to get together and hold hands before launch, but it turned out to be the right call. It's a testament to how conservative we are at JPL. Once we got to the rocks, they were very soft, so I don't think there'll be an issue within the drill's lifetime.”

  From your mouth to Mars's ears, my friend.

  By the end of 2012, Curiosity had descended into Yellowknife Bay proper and was looking for a rock to drill into. Of course, there were plenty around—everywhere you looked, in fact. But it had to be big enough to sit still for the drilling, and preferably the first target should not be too hard—nobody wanted to break the drill on the first try. And, of course, the rock should be positioned in such a way, and of the proper composition, to yield a maximum science return.

  Like everything else in this mission, patience was a virtue. John Grotzinger meant what he said after the whole “history books” debacle.

  For the press, the news conferences at JPL were mostly over, at least ones that we could attend. Most were now being held by teleconference, which is a great time-saver. JPL and NASA have the system pretty well nailed down, and I was eagerly anticipating the results of the drilling along with tens of thousands of others.

  Yellowknife Bay was a shallow depression past Glenelg and Rocknest. It sounds like the rover is really driving over hill and dale to all these places with funny names. But they are not quite what we would normally think of as distinct “places.” The distance from Glenelg to Yellowknife Bay is about the same as from your back fence to the front door of your house—something in the neighborhood of two hundred feet. It's a bit like naming the rocks, shrubs, and gopher holes in your yard. But at the speed that Curiosity was driving, and given the extreme and rich geological diversity of the part of Mars it was traversing, it makes sense to create names on a map. For one thing, it helps the drivers and science teams (not to mention the rest of us) keep track of where Curiosity is. It also makes Mars feel friendlier. It's not much fun to move from rock N165 (the numeric designator for Coronation) to 4.59°S by 137.44°E (the positional measurement for Hottah).

  As the rover moved into Yellowknife, the geologists started looking for a drilling target. They had lasered rocks, sniffed boulders, and tested the sandy soil of Rocknest. The atmosphere had been analyzed. The relevant instruments, SAM and CheMin, had been commissioned and tested. The last major item on the mechanical list was the drill.

  As the rover made the shallow descent into the “bay” (just a few feet of gradual drop, really), the Mastcam and ChemCam were busy looking at potential targets. The wonder of the laser-driven ChemCam was becoming apparent now; what could have easily taken days or even weeks took only hours. With the ability to shoot the laser and image the resulting spectra from a distance, the device saved not only drive time but also the long and careful process of bringing the arm down and rotating the APXS into place to examine a rock.

  One more tool that needed to be tested was the wire brush, the Dust Removal Tool (or DRT). The rotary wire bristle was capable of scraping the dust off of rocks for a better look. This too would help them in the all-important job of finding a rock to drill.

  The wire brush was used on a rock they named “Ekwir_1.” Diana Trujillo, lead for the brush team (everything on Curiosity has a team) said in a NASA interview, “We wanted to be sure we had an optimal target for the first use.” She added, “we need to place the instrument within less than half an inch of the target without putting the hardware at risk.” As we heard in a previous chapter, they even obsess about how long and how hard to press the DRT against the rock in order to avoid bending the wires on the brush.

  And they even had to take the ambient temperature into consideration. Anything on the arm that actually touched a Mars rock, or even the soil, was at potential risk of breakage, or causing arm damage, due to any forces imparted from the target object through the instrument and into the arm. The rover could always slip, but they were careful in these early efforts to pick flat areas on which to park Curiosity while running tests. But if the temperature shifted too much while they were working, even the small expansion or contraction of the arm could be a problem. Of course, this was factored into the simulator, but let's look at a theoretical example. Say they were using the drill on a rock. Something snags, or the rover has a software hiccup that freezes the arm overnight. The temperature differential between night and day in Gale Crater, at the time, was almost 120°F. That broad temperature swing would, as it got colder, cause the m
etal in the arm to contract—and much more than you might think. As the arm contracted, if the drill was sitting in a borehole, it could at the least bend or even break the drill bit. At worst, it could destroy the drill apparatus or damage the arm servos. Then, as the sun rose, the arm would heat and expand. The combined forces could make for a bad morning.

  So with either the drill or the brush, nobody wanted the arm moving in ways that were not planned for. So they picked a nice, flat rock on a nice, flat surface to work on. Once the DRT instrument had done its work, the MAHLI camera was used to take a closer look. The resulting images were something to warm the hearts of the geologists. The rock went from the usual ruddy, oxidized red to a grayer shade, and pits and veins could be seen. It was not remarkable, but it was a happy day, as the brush worked and they could move on to find a drill site.

  They soon had their man, so to speak. A flat, veined rock the geologists named John Klein was selected. The namesake John Klein had been the deputy project manager of MSL until 2009, and died unexpectedly six months before launch. It seemed a fitting tribute. Curiosity slowly rolled to the rock.

  Part of what attracted them to Klein and the area in which it dwelled was the oft-cited property of thermal inertia. Orbital images with infrared cameras had shown that this area cooled more slowly in the Martian night than other surrounding terrain. The slower a rock cools, the denser it is. And dense rocks in a lake-bed environment could mean sedimentation, and the denser the sedimentation, the older it might be and the more information it might hold. As Grotzinger put it, “The orbital signal drew us here, but what we found when we arrived has been a great surprise…this area had a different type of wet environment than the streambed where we landed, maybe a few different types of wet environments.”