by Rod Pyle
The landing software included a provision for a gliding entry. The fifteen-foot heat shield was based on the old Viking design. It was not a pure cone, but had an offset bulge on the bottom, making it, in effect, a low-efficiency glider. As it sped toward the surface, it was able to “surf” on the thin blanket of air, scrubbing off some of its blinding speed and allowing it to steer toward the landing area. The maneuvering thrusters helped to keep it aimed true. There were also eight tungsten weights, heavy gray cylinders, mounted inside the edge of the structure.
Fig. 15.2. SEVEN MINUTES OF TERROR: This chart shows each phase of the entry, descent, and landing (EDL) sequence. Top left is atmospheric entry, bottom right is landing. Image from NASA/JPL-Caltech.
As the heat shield began to glow white hot, it burned away, a process called ablating. This is how most heat shields work—heat is carried away by the burning materials as they are shed. The trick is to make sure that it burns evenly, so that it will maintain its aerodynamic shape.
The math needed to glide properly was based on work done well over fifty years earlier for the Apollo program. The capsules returning from the moon also needed to glide, or, more precisely in their case, skip along the atmosphere as they came home to reduce speed. While Martian air is much thinner than Earth's, the same principles applied.
To glide effectively, the spacecraft had to shift its center of gravity. This was accomplished by ejecting the first two weights, each 170 pounds. The craft pivoted to aim the offset heat shield in the proper orientation for gliding and steering. Fine adjustment was accomplished by the small rocket thrusters.
By the time this portion of the atmospheric entry was complete, MSL had slowed to Mach 1.7, or about 1,300 mph. It had lost a lot of speed from the roughly 13,000 mph it was traveling when it reached Mars, but it was still going way too fast. Six more tungsten weights, each weighing fifty-five pounds, were then tossed away, bringing the spacecraft back to an even keel. At just over six miles from the surface, the parachute deployed. The parachute had given the engineers fits, as it had ripped repeatedly during testing. And working with cloth is a nightmare in these kinds of tests. Metal is nice and predictable—bend or pull it this much, and it will weaken, this much more, and it will break. Cloth becomes a somewhat-unpredictable mass of small fibers that is hard to quantify, and some people were still nervous about the parachute's performance. For Martian landings, only Viking's parachute had been of a similar size, but that spacecraft was only about half as heavy as MSL. There were just so damn many variables. But they had tested the fifty-foot chute in the nation's largest wind tunnel, and it had held at speeds in excess of this, so it should work….
It did. The parachute deployed, then widened as it un-reefed (a gradual releasing to full width), slowing the lander. As it settled into a mostly vertical fall, the heat shield, its work done, was dropped. The MARDI camera started to record the final phase of the descent, and would provide dramatic imagery of the ever-nearing landscape. It was then that the Mars Reconnaissance Orbiter tried for a picture of MSL during its final few minutes in the air, aiming at what the MRO controllers had predicted would be the right place. Amazingly, it worked, and provided a stunning photo of a beautiful, billowing parachute.
Two minutes to go. Flight controllers back on Earth were still monitoring signals that indicated that the spacecraft was outside the atmosphere due to the long delay for the radio signals in reaching Earth. MSL was way ahead of them, flying on its own.
With just over a mile to traverse and still traveling nearly 200 mph, the rover and its rocket pack dropped free from the remainder of the aeroshell, leaving it and the parachute behind. This was the terminal phase of the landing, and the scariest for its creators.
It was time for sky crane to go to work.
“What the f—” were my timeless first words upon encountering sky crane. That was my inner voice. The outer voice muttered something like, “By golly, that's interesting.” I would later learn that many in the planetary-science community, and beyond it, uttered something more akin to the former than the latter. The approach was weird.
As a person who had written about, covered, and tracked the US space program for a couple of decades, this struck me as a decidedly different approach to landing on another world. Viking had landed like a proper, macho lander—upright, rockets blazing after a parachute-slowed entry into the Martian atmosphere. With Pathfinder, things began to get weird—the second US mission to land on Mars did so covered in fabric beach balls, bouncing dozens of feet into the thin Martian air before rolling to an eventual stop. The Mars Exploration Rovers, Spirit and Opportunity, arrived in a similar, somewhat–Bugs Bunny style, bouncing over a dozen times, then rolling to a stop as their airbags deflated.
And then there was the Mars Science Laboratory. How the heck do you land something that large, and heavy, and that is traveling at interplanetary speeds?
“It was thinking out of the box. In fact, we threw away the box. We were literally going through all possible ways to land this machine, trying to imagine every possible configuration, whether it made sense or not. You do this until you bump into one that's so crazy that it makes sense.” That's Rob Manning once more. Again, he was at the nexus of figuring out how to set down a machine on Mars while staying within budget. Again he led a small team of equally brilliant people to work through the problem. And again, they came up with something that made people sit up, swallow hard, and strain their minds to get their head wrapped around it.
I realize that I left you hanging a few dozen feet above Mars in the last chapter. But before we look deeper into how sky crane works, I want to share its origins. Nobody can do that in a more entertaining way than Manning, so let's hear the rest of the tale from him.
Fig. 16.1. MSL'S PIECES: This graphic shows each component of MSL: (1) the cruise stage (discarded before atmospheric entry), (2) the aeroshell that protected the rover and descent stage during entry, (3) the decent stage/rocket pack, (4) the rover, Curiosity, (5) the heat shield, and (6) the parachute. Image from NASA/JPL-Caltech.
“So NASA had just experienced two embarrassing failures with the Mars Polar Lander and the Mars Climate Orbiter in 1999. Before these missions failed, I was leading the effort to build an oversized version of the Mars Polar Lander to carry a rocket and a sample-return rover on the top deck. It was supposed to land in 2003 and 2005, there were going to be two of them….”
As you will recall, the Mars Polar Lander and Mars Climate Orbiter missions were both part of the “faster, better, cheaper” era. And, unlike Pathfinder, both failed. It was clear to NASA and JPL that the next few missions had better work. The MER rovers landed just fine, but the landing system for the big MSL mission was being discussed before MER's successes, and the engineers were understandably a bit gun-shy.
Back now to Rob and his big, oversized post–Mars Polar Lander mission: “So these [new] landers were supposed to collect samples, rocket them off towards Earth, and by 2011 we would have samples from Mars. But we previously hadn't spent a lot of time trying to figure out if something that size and that scale would work, so my team at Lockheed was struggling with it. We were trying to make a payload [rover] that could drive off the top deck of the lander.” Remember that Pathfinder and MER had used landers that folded open, the sides becoming ramps for the rovers. That worked well for smaller machines, but you start getting up into the big leagues—say, a bulky two thousand pounds or so—and this approach does not scale up well.
Rob continues: “The ramps were getting huge, the rover was getting huge, and the lander was so big by now we realized we did not have rock clearance.” You need to be able to land on rocky surfaces when sending machines to the Martian surface, and at this scale they were having trouble guaranteeing clearance of larger boulders. “Trying to find a place on Mars that is rock free is still a big challenge.” Indeed, it's like trying to find a part of the Atlantic that is dry. Using the vastly improved images from Mike Malin's orbiting cameras, howeve
r, they were indeed able to find areas that looked fairly safe. But that presented a new problem….“The geologists want you to go to rocky places so that they can investigate the rocks, so to accommodate them we were struggling to get that big lander down safely. Then these two failures happened, and we said to ourselves ‘Okay, we have some serious issues here getting things to Mars safely, irrespective of why Mars Polar Lander disappeared. Let's go back and rethink landing on Mars.’ So we did.”
Two failures in a row is enough to make even the most enthusiastic, optimistic engineer a bit cautious. It was time to start with a clean sheet of paper. The traditional approach to landing a big, heavy rover was, at the time, very much like the Viking lander—use a large descent stage that lands on legs with rockets firing. The problem was not so much the lander, though, it was the rover—putting a one-ton rover on top of a lander creates problems. The dynamics get very wonky, and it's a bit like trying to balance a bowling ball on a broomstick—the center of gravity is very high and the system becomes unstable. So they looked at all the ways they could think of to get around the balance problem.
“A small group of us got together, and we fiddled with all these permutations,” says Manning. “One of them was a variation on Mars Pathfinder. We would have to replace the solid rocket motors with liquid rockets that could be throttled. Then we would lower the lander on ropes, and we wouldn't need big airbags anymore—if you control it well enough, you landed using the ropes with no airbags at all. You just land your rover right on its wheels.” This did away with the issues of a rover sitting on top of a landing stage, but was, to say the least, an unorthodox approach. Manning remembers his first feedback on the idea: “The three of us were pretty excited about this. We started talking to some of the dynamics-control guys [the engineers involved in flying the lander to the surface of Mars] and they said, ‘Look—we just don't know enough about the dynamics of landing a system this way…we can study it later, but let's not do it now—it's just too crazy.’ So we put that idea off to one side.” What the engineers said when Manning and crew left the room is not recorded….
Sky crane was sidelined, but not for long.
Manning continued: “MER had not happened at that point. The whole idea of landing a rover on its wheels with the control system overhead, suspended by ropes seemed a little out there. So we had another idea which looked more promising called the pallet lander. It's a version of the lander that's more crushable.”
When any machine lands on another planet, the challenge is to lose energy, in this case downward speed, before or when you touch the ground. If you have lots of money and big, powerful rockets, you can do it the way Viking did—slow yourself on those big, expensive rockets until you make ground contact, on shock-absorbing legs. Or you can use other ways to slow down, to absorb the shock. The pallet system would allow the whole landing stage to crush itself in a controlled fashion, which would result in the absorption of a lot of energy (and, in effect, speed reduction) and get the top deck, and the rover, closer to the ground.
“So the platform lands and crushes itself. It is very stable, and you can actually drive off this flattened lander, which is more conforming to the surface.” Again, thinking outside the box—in effect, destroying the spacecraft's lower stage as it lands to give the rover a knee-level roll-off to the ground. Brilliant. “We studied that for a couple of years. At that point the only lander mission we had put together besides Viking was Pathfinder. Mars Polar Lander was dead and the Mars Exploration Rover mission hadn't started.”
Manning was moved over to the MER program to make his airbags and rovers work. But MSL marched on, says Manning, and a new factor entered the discussion—those bigger, variable engines we talked about that landed Viking so successfully. “A very small team was still looking at ways to make the crushable-pallet system work. That's when we got money to develop the throttleable engines.”
Manning may have moved to MER, but the heavier, more capable MSL rover was always in the back of his mind—lurking, waiting, taunting. He continued: “We spent three years getting MER off the ground and to Mars, but while we were developing MER, I realized that we were learning a lot about the dynamics of using rope up on Mars. That idea that we had—the precursor to sky crane—is not as crazy as people thought.” It was time to campaign a bit. “I tried to talk to the current manager of MSL—which was then called the Mars Smart Lander—and he was interested. They got some workshops together and started having discussions about this option.”
The workshops accomplished what they were intended to do—collect a bunch of smart people, look over the problem from all angles, then assault each other's thinking until you're pretty sure it can be done the way you're discussing. Remember the “hazing” John Grotzinger discussed regarding sedimentation on Mars? This was similar. “So after a few workshops between 2000 and 2003, we decided to should look into this approach more carefully.” The MSL team looked at the dynamics of a rover slung below the lander. “Since many of us were still busy on MER, we would come swooping in and then talk about the MSL landing system, then go back over to MER.” It was like drive-by brainstorming—spend a few hours on the Opportunity and Spirit rovers, sneak off to meetings or working sessions about MSL, then get back to the MER project. But soon, MER landed and there was more time to consider options for MSL. They decided that the nascent sky-crane system was their best option. “Now the hard part was convincing the rest of the world,” Manning says with a wry grin.
“There were a lot of levels to convince. The next was JPL upper management, then the outside world. So we brought in a lot of independent people.” These included experts from a McDonnell Douglas rocket project from the 1990s called the DCX, Sikorsky Helicopters, people who worked on the Viking and Apollo programs, and many others. “We even had Harrison Schmitt, an astronaut from Apollo 17,” Manning remembers. “So we were going through the design, and everyone is poking at it, poking at it, and poking at it. We then took their recommendations and went back and worked on it for about six months.”
They had what could be politely called “nagging doubt” to overcome. Manning calls it the “laugh test.” “Their first response early on was: ‘Are these guys serious? They seem serious!’ The biggest question was, why would you go to something so outrageous if you have all these other ways you've done it successfully before?” The reviewers all knew about Viking, Pathfinder, and MER, so they understood landing with rockets on legs and, having overcome nagging doubt once before, the bouncing airbag method. But now…this? “So we had to explain for quite some time what the weaknesses were for all these other approaches. Sky crane was the only one that did not seem to have an Achilles heel. It seemed to work…at least on paper.”
The design was indeed fascinating on paper, but it left a lot of problems to be worked out. First they needed a way to slow the spacecraft to a crawl before sky crane could even begin to do its job. That required parachutes and rockets. Once they had slowed their descent to a (fast) walking pace, “We needed to communicate between the rover and the descent stage, and we needed a way to lower the rover down.”
That's where the ropes he spoke of earlier came in. And the ropes are what engendered most of the horrified looks—or giggles—from onlookers.
“At this point we were a bunch of true believers,” he recalls with conviction. “We needed not just management but outsiders who really knew the area to believe in it as well. The hard part now was going all the way to the top,” meaning NASA senior management. Here they got lucky. The NASA administrator at the time was Michael Griffin. “Mike was a quintessential aerospace engineer, probably the most overqualified NASA administrator and history in that light. He's got more sheepskin than any other. So once we got to that level, the only person we really had to convince was him—Mike Griffin.” Manning is clearly fond of the man. “He's not a delegator, he's a ‘show me’ kind of person. The way he worked was, ‘If you can convince me, this is going to happen. If you can't convince me, th
en it's not going to happen.’” Adam Steltzner, who had taken over management of the entry, descent, and landing part of MSL, gave the presentation at NASA headquarters in Washington, DC. Griffin listened coolly at first, then with increasing interest. Manning continues: “Griffin finally said, ‘Well if anybody can do this, you can.’ I'm pretty sure he was thinking this is not how he would land on Mars; he might even have been thinking it was a bit Buck Rogers. But he let us do it. I'm sure Cadwell Johnson [of the original Pathfinder review group] would've said, ‘Hey, it's supposed to have four legs you jerks!’”
So they had NASA's top vote in hand. Of course, at the same time that this top vote cleared the way to move ahead, it also meant a lot of pressure—the administrator, NASA's de facto CEO, is watching and has a personal investment in the project. And, he's a senior aerospace engineer—there's nowhere to hide.
They had permission. They had a budget—for now. Now they had to make it work.
I'm not an engineer. I don't have an advanced degree in the hard sciences. And when I saw sky crane, just a few years after JPL's two Mars missions had augered-in back in 1999, I was stunned. There appeared to be an awful lot of potential points of failure in this system—many dozens of explosive bolts (pyrotechnic fasteners that release when they explode), any one of which could decide not to work. There were the throttleable rockets on the backpack—wasn't Viking the last time they had built those? It also had to navigate to a very small zone in the middle of a deep crater—and do so mostly autonomously. And then there were those damn ropes.
