The One-Pound Problem

All the Mars Ascent Vehicle has to do is deliver 16 ounces of rocks in a container the size of a grapefruit to Martian orbit. If only it were as easy as it sounds.

When Brian Wilcox was growing up in the 1950s and '60s, his father was in the space business, and for a kid whose hobby was model rocketry, what could be better than that? Howard Wilcox worked on early launch systems for the Naval Ordnance Test Station in China Lake, California, then later for General Motors, where he designed robots that could have scouted landing sites for the Apollo astronauts (but never did). In those visionary days, when the rules of spaceflight were still being written, Brian and his father would have long talks about what kind of launch vehicles we would someday need on Mars, and whether the small model rockets he and his friends built could--in principle--reach Earth orbit.

The answer, sadly, was no. A former physics professor at Berkeley, Howard Wilcox knew it wasn't gravity that dictates the minimum size for a launcher. "My father explained to me at a very young age that the only limitation was the atmosphere," Brian recalls. "If you didn't have an atmosphere, you could make them small." A terrestrial rocket has to push through a plug of air equivalent to a 30-foot column of water, and physics dictates that the smallest vehicle capable of moving all that atmospheric mass without paying a penalty in momentum is about 30 feet long.

On Mars, though, the thin atmosphere is equivalent to only about four inches of water. Many years later, working as an engineer at NASA's Jet Propulsion Laboratory, Brian Wilcox would participate in a key series of workshops to brainstorm ideas for a Mars sample-return mission, the most ambitious planetary project in history. One of the critical questions facing the workshop attendees, he remembers, was: What was the smallest rocket that could possibly leave the surface of Mars and make it into space? "And I knew that answer because I'd discussed it with my father many times. The answer is: about the size of a pencil."

Howard Wilcox didn't live to see his son's triumph on July 4, 1997, when, as a member of JPL's Mars Pathfinder team, Brian helped pull off the first spacecraft landing on Mars in more than two decades. Pathfinder was a watershed event for NASA. Its $265 million price tag proved that planetary missions didn't have to be expensive to be fun, and it gave planners confidence that the agency's goal to return samples of Martian soil to Earth by the middle of the next decade--which had just received a huge boost from the discovery of signs of possible fossil life in a Martian meteorite (see "Pieces of the Rock," Apr./May 1997)--was feasible for around $1 billion.

By the following spring, however, the confidence had all but evaporated. The sample-return mission--in fact the whole Mars exploration "architecture," which called for launching at least one spacecraft to the planet at every two-year opportunity--was running into severe financial and technical problems and was on the verge of meltdown.

The plan at the time of the Pathfinder landing called for three missions directed toward the goal of getting samples back from Mars. The first two would send large, well-equipped rovers to Mars in 2001 and 2003 to dig and drill rock samples from different sites, then "cache" them in sealed containers on the surface to await pickup. The third would dispatch a smaller "fetch rover" in 2005 to whichever of the two caches looked more promising scientifically. The little rover would trundle over to the chosen sample, bring it back to the lander, and transfer it to a rocket called the Mars Ascent Vehicle (MAV), which would send it back to Earth.

The scheme quickly ran into difficulties, though, that threatened to break the project's budget and schedule. The package launched to Mars in 2005--which included the lander, the fetch rover, the rocket, and miscellaneous other equipment--would have to fit inside a medium-size Delta launcher to make the mission affordable. And designing a lightweight Mars rocket that could live within that limit was turning into a very tough challenge.

The MAV was to be liquid-fueled, so it needed exotic propellant that wouldn't freeze during the Martian night. It required fancy plumbing, miniaturized components, and lightweight motors and fuel tanks. "It needed a whole bunch of stuff to try and get this thing down to a low mass," recalls Mark Adler, chief engineer for the sample return mission. "That was all extremely expensive." The price for the MAV ballooned to $120 million at a time when the entire project was budgeted for only $200 million a year. And, at 1,300 pounds, the thing was way overweight.

The MAV wasn't the only headache. The big sample-collecting rover was having its own development troubles, and getting it ready by 2001 looked like a long shot. Meanwhile, the human exploration office at NASA, realizing that sample return was the only train leaving for Mars, proposed adding an experiment to the 2001 mission that would benefit future astronauts: a prototype propellant plant to extract oxygen and hydrogen from the Martian environment and make fuel for the rocket. The only catch was that the money had to come from JPL's fixed budget, which was now strained to the breaking point.

By the time Bill O'Neill, a manager who had come from the Galileo Jupiter mission, took over as head of the sample-return project early last year, "a lot of issues were out of control," he says. "As things unraveled, all the senior managers kind of pitched in and said, 'We've got to figure out a way to fix this.' " One of the things O'Neill did was hold a pair of workshops in spring 1998 at the Embassy Suites hotel in Arcadia, just down the freeway from JPL. His notion was to invite fresh ideas from inside and outside the team on how best to do a sample return.

Brian Wilcox wasn't working on the project--his expertise is rover design--and he was a bit self-conscious venturing outside his turf. But he asked to present a concept for a Mars rocket he'd had in the back of his mind for years, based on work his father had done at China Lake in the 1950s. "I was thrilled to be able to step forward and say 'Hey, I've got another way. And by the way, I've got a 19-minute videotape that shows how it can be done.' "

The video shows Howard Wilcox sporting a bow tie and crewcut, standing at a blackboard circa 1958, and looking for all the world like Mister Wizard as he chalks out the basics of a then-secret project called NOTSNIK. Wilcox and his team attempted--some think they succeeded in--launching a two-pound satellite to Earth orbit from under the wing of a Douglas Skyray fighter (see "The China Lake Launches," Feb./Mar. 1997). One of their innovations was to launch NOTSNIK's last rocket stage, the one attached to the satellite, with its nozzle pointing in a direction opposite from the other stages. When it reached space it coasted for half an orbit, spinning all the while to maintain its orientation, so that on the opposite side of Earth its "backward" nozzle was pointing in the right direction to boost the satellite to its final orbit. (Think of the spacecraft as a car on a ferris wheel, its nozzle pointed down on its downward arc. With its attitude held stable, its nozzle will continue to point down as the craft begins to climb and can accelerate it in the direction of its forward motion.) It was a clever way to avoid putting guidance electronics on the last stage and save precious weight in a system that measured performance in ounces.

What Wilcox proposed to the JPL workshop wasn't the theoretical pencil-size Mars rocket he and his father had once dreamed up, but it was close. The "Mini-MAV," as it was dubbed, weighed a mere 40 pounds and stood only three feet high. It could deliver seven ounces--200 grams--of Martian rock to orbit, a significant step toward the payload the team now hopes for. The first stage would be set spinning rapidly on a turntable just before launch, the second stage would spin up with small thrusters, and the third stage would pull its little NOTSNIK trick to point itself in the right direction. No active guidance at all. The Mini-MAV wouldn't need it, because another key assumption was changing at around the same time. Sample caching would now be done in Martian orbit instead of on the surface. A dumb--that is, imprecise--rocket could just lob the sample canister into space, where a smart orbiting spacecraft--supplied by France and sent to Mars by the powerful Ariane 5--would find it, rendezvous with it, capture the canister, and send it back to Earth.

Scrapping the guidance system was "a very big deal," says Mark Adler, because it alone weighed tens of pounds compared to less than a pound for the sample itself. Any mass removed from the top of the rocket had tremendous leverage on the whole system. The ratio was eight to one: For every pound cut from the payload, you'd save eight in motor and fuel at the bottom.

At first, Mini-MAV appeared to be one of those rare engineering miracles that offers nothing but advantages. It was so lightweight that you could now get away with using lower-performance solid-rocket engines instead of liquid engines, with all their fussy plumbing and sensitivity to the cold. And solid rockets, with their relatively simple technology, could be ready by 2003, while the large liquid MAV couldn't. Finally, NASA could return samples from both cached sites, one in 2003 and one in 2005.

That summer, Wilcox gave presentations on his idea to other groups hashing out the Mars architecture, and no one shot it down. In fact, with each level of review the enthusiasm only mounted. One veteran space consultant whom O'Neill had hired to help rethink the sample-return mission was so jazzed about Mini-MAV that he canceled a long-planned European vacation to keep working on it. Adler was excited too, but cautious. "My initial thought was Gee, that's a good idea. I wonder if it works." A couple of small teams at JPL were asked to give it a closer look.

Doug Caldwell, a young engineer who had left the computer industry six years earlier to join JPL, took over as the new head of the MAV office in September 1998. Before that, working on the Deep Space 1 technology demo mission, he had heard "a crazy idea of a spinning rocket that will weigh only 30 or 40 kilograms," and was well aware of the buzz surrounding it. Outside JPL, scientists and NASA managers who the year before had watched in horror as the sample-return mission nearly collapsed were now talking up Mini-MAV "like it's going to completely save the world," says Caldwell. And he thought to himself, Yeah, but you don't actually have to build it!

He thought the idea was cool too, and still does. But he knew that "whenever someone throws out something brand new, it almost always looks better" than what you've got in front of you. And the study teams were already beginning to find cracks in the Mini-MAV miracle. "A lot of the simplifications turned out to be oversimplifications," says Caldwell.

For starters, the rocket's conjectured 45-pound weight hadn't included the turntable on which it rested, the "igloo" covering needed to keep it warm, or other sundry equipment not on the rocket itself. Add another 35 pounds right off the bat.

The NOTSNIK had done without a guidance system partly by using fins to stabilize itself in the atmosphere, but the Martian air was too thin to help here. Plus, the NOTSNIK engineers hadn't been picky about where their satellite wound up, as long as it made it into orbit. The Mini-MAV couldn't be quite that loose--it had to come within striking distance of the Mars orbiter.

"God created guidance systems for a purpose," says Caldwell. So the engineers admitted defeat and added active control--thrusters, inertial position sensors, computers, and power--to the first stage, which pushed the weight up from 80 pounds to almost 300.

Another issue the Mini-MAV design hadn't addressed was how the sample canister would keep the Martian dirt pristine. This was important not only to scientists, who wanted uncontaminated samples, but to NASA's "planetary protection" watchdogs, who had to guarantee to the public that no dangerous Martian bug--a remote possibility, but not entirely out of the question--would be returned to Earth. The sample-return project hadn't yet come to grips with the planetary protection dilemma when Brian Wilcox proposed his Mini-MAV. By the time it did, the sphere that could hold the Martian dirt and keep it sterile weighed ten times more than the sample itself.

This creeping weight gain was fairly typical for the early design phase of a space engineering project, says Caldwell. "Push the system here and it bulges out there with some kind of problem--too much heat, or cost, or weight." It was mainly that damn eight-to-one mass penalty. When anyone suggested adding another pound or two to the sample canister, he says, "I wanted them to know how painful it is."

By the spring of this year, the Mini-MAV had ballooned up to a 375-pound, plain old MAV, which stood about as tall as a person. It was still a lot better than the old liquid-fueled MAV, which had trimmed down to 600 pounds but appeared to be stuck there. Even so, with all the commotion that had been made over the Mini-MAV, Caldwell couldn't help feeling like the guy who's been given a perfect handoff, only to fumble it inches from the goal line.

Asked what's the hardest part of bringing back a sample from Mars, Bill O'Neill thinks for a moment, looks down at his desk, and takes in a deep breath. Then he lets it out with a laugh. "It's all hard!"

He considers each task of the three missions in order (see diagram, opposite page). Getting to Mars is straightforward enough. But two of the flight plans call for "direct insertion" by aerocapture. In other words, the spacecraft will come screaming into the Martian atmosphere, where drag, not retrorockets, will slow it down. It's never been tried.

Once safely on the ground, the 2003 and 2005 landers will deploy rovers about the size of a child's wagon, which will roam the Martian surface, collecting rocks and drilling cores from some 20 sites over the course of three months. The most interesting samples will go into a cache box, which will be transferred to the MAV when the rover returns to the lander. A mechanical arm then hoists the MAV to vertical launch position, and off it goes.

Still pondering the relative risks, O'Neill says the landing on Mars isn't too bad, since it's been done before. The rover can be tested on Earth. The transfer from rover to MAV is tricky, but also can be practiced beforehand. "I suppose if I had to pick something, I would say the ascent vehicle, because it's the least demonstratable," says O'Neill. "There's no environment you can find on Earth that can completely mimic what we're going to do at Mars."

The sample-return strategists will therefore rely mostly on computer simulation, just like the Mars Pathfinder team did, testing only those few key elements that are particularly worrisome and are possible to try out on Earth. For example, Doug Caldwell wants to run what he calls a "burp test" in a 100-foot-high chamber filled with simulated Martian atmosphere. A MAV would be loaded with just enough solid propellant to burn for a second or so--long enough for the rocket to clear the launch deck so the engineers can see if its exhaust plume has some funny interaction with the lander.

The sample return mission becomes much more difficult the more Martian dirt the project is asked to bring back, and last spring this was a topic of vigorous discussion between JPL and NASA Headquarters. By then Caldwell and company knew how to build a MAV that could lift about 12 ounces--350 grams. But NASA was "adamant," says O'Neill, that it wanted the full pound--a little more than a pound, in fact, 500 grams, about 17 ounces--and was pushing for a "stretch goal" of 1,000 grams.

The can-do team at JPL didn't argue that it was impossible. "We wouldn't want to immediately take the easy way," says O'Neill. But bringing back more sample doesn't automatically increase the scientific productivity of the mission. The project had long ago rejected what was known as the "grab sample" scenario--dash to Mars, dig up a bunch of dirt in a hurry from right around the lander, then rush it back to Earth. To maximize their chance of finding evidence of water or some other geologic prize, the scientists much prefer collecting smaller samples from many diverse sites, with cameras and other instruments on the rover carefully documenting each setting. All they need is a few milligrams back in the lab--if it's the right rock.

This requires that the cache box be divided like a honeycomb into compartments, each containing a sealed sample from a different site. The greatest worry is time. It takes time for scientists on Earth to study the surface photos and determine which sites they want to explore, time for the rover to move on to the next location, time to photograph the site up close, time to inspect each rock in the vicinity, time to drill the inch-long cores, and time to seal them in the compartments. The rover is expected to last only 90 days on the surface before dust and general wear and tear reduce the effectiveness of its solar arrays and it runs out of electrical power.

"Our job would be a lot easier if we could power the landers with nuclear batteries the way Viking did," says O'Neill. "But it's politically no longer acceptable." So three months is all the rover will have to do its job, and some question whether that's long enough to collect 500 grams of material from 20 different sites. "We'll be hard pressed to get enough stuff to fill up the cache boxes," says O'Neill. Steven Squyres of Cornell University, one of the lead scientists for the sample return, shares his concern. "This mission is going to be like that old supermarket sweepstakes, where you've got 90 seconds to go through the aisles, grabbing everything you can," he says.

Returning more than 500 grams could mean that some of the stuff gets tossed quickly into an undifferentiated bin at the end of the rover's lifetime, which isn't as appealing to scientists. And this loose dirt would have to be balanced in some way so it didn't slosh around when the MAV lifted off.

That, plus the general penalty for adding mass to the sample return canister, made even a few extra grams of material worth arguing about. JPL now feels comfortable with the 500-gram requirement, but "We don't know how we could get from 500 to 1,000 grams," says O'Neill. The difference is about the weight of a slim hardbound book.

On an unseasonably cold April day in Pasadena, with patches of snow still clinging to the peaks of the San Gabriel mountains that abut the JPL campus, around 50 engineers and technical managers from the space industry gathered, at Caldwell's invitation, for an all-day briefing on MAV. In coats and ties, unusual attire for JPL, they sat mostly silent while O'Neill tried to get them as pumped up as government contractors are allowed to get.

"The first-ever launch from another planet--what a great thing to be able to participate in!" he enthused, kicking off the meeting with a quick Vu-Graph walk-through of the sample-return mission. Bringing back a piece of another world, he continued, was "one of the few remaining firsts in planetary exploration," a historic endeavor on a par with Sputnik or Apollo.

Caldwell, introducing himself as the "MAVman," explained the reason for the meeting: JPL wanted an outside opinion. Most of the $60 million to be spent on the rocket system would go to contractors once actual fabrication began. Right now, though, he was looking for a reality check before going any further. "Do we have the right architecture?" he asked.

The MAV design was holding at about 375 pounds, and mission planners were hoping to drop to 350. It still called for three solid rocket stages: a guided first stage, spinning second and third stages, plus that additional NOTSNIK back-end-first approach on the third. One area in which the team needed advice was on the fuel composition for the solid rockets, because spinning had certain drawbacks. A key reason for rejecting the spinning first stage early on was that it required something like 500 rpm for accurate pointing. But even with a slower spin rate for the upper stages, people were concerned about the buildup of aluminum slag. Solid rockets on Earth typically are made with 16 to 18 percent aluminum. The more aluminum, the more oomph. But when the fuel burns, it produces aluminum oxide slag, which in a spinning rocket would throw off the balance like an uneven load of laundry in a washing machine. The answer appeared to be to cut the aluminum content way back, to one or two percent, but the exact formulation would have to be determined. Caldwell wanted a new team to look just at propulsion.

Some of the people attending the April briefing worked for companies--notably Lockheed Martin, part of which used to be Martin Marietta--that had been fiddling with Mars rocket concepts for at least 20 years. Most of these were old-style behemoth designs unsuited to the more economy-minded NASA of the 1990s. Later, back in his office, Mark Adler pointed to one of the old artist's conceptions showing a big, heavy rocket with an elaborate gantry--on Mars. "This is part of a ten-billion-dollar sample-return mission that never flew," he chuckled. "I can't imagine why not."

Still, there were some good, experienced Mars mission designers in industry, and by mid-summer Caldwell had hired two study teams--one to look at propulsion and the other to evaluate the overall system architecture with a fresh perspective. Almost immediately the teams found ways to improve the MAV. Based on some of the contractor presentations, it started to look like two stages would be better than three. It was a lot simpler and it saved weight: one less interface between stages, no third propulsion system. Even without guidance on the second and third stages, the trade studies showed that the old version was still heavier and more expensive. It was actually pretty obvious, says Caldwell. But "you get mentally tied to a concept," and the JPL engineers had been trying their damnedest to make the MAV work with the spinning upper stages stabilized NOTSNIK-style. Now none of that would be necessary.

What had Brian Wilcox's novel idea, so celebrated just a few months earlier, really contributed then? "What it really did was force the thinking in a different direction," says Caldwell--away from big, heavy liquid engines and a design gridlock that had all but paralyzed the project.

Frank Jordan, a senior JPL manager in charge of Mars mission planning and an old hand at planetary spacecraft design, thinks that Mini-MAV deserved its accolades. "Most of the rah-rah was justified, in my view," he says. "I mean, it's not a 20-kilogram little-bitty rocket anymore, but it's 170 kilograms, and that's better than 300 kilograms." Jordan has seen a lot of engineering concepts come and go. He worked on the first mission to orbit Mars, Mariner 9, which entered Martian orbit in 1971. "So much of engineering, I think, works this way," he says. "You have some really innovative something, and by the time you bring it to fruition it's compromised--but still better."

The Mars rocket is still evolving, and it's still hard work. Caldwell and his team hope to have the design pretty well nailed down by December. Meanwhile, Wilcox has moved on to other projects and, like most people at JPL, has several things on his plate. He doesn't especially mind that in the end, his Mini-MAV didn't prove practical for the Mars sample return. Logic prevailed just as it did in the 1950s, when his father was trying to orbit another small object around another planet. Back then, scrapping the guidance system made sense because it weighed a lot more than NOTSNIK's payload. Once the payload grows beyond a certain size, the advantage is always with active guidance, and the MAV sample canister turned out to be much heavier than Mini-MAV's 200-gram payload.

Wilcox spends most of his time these days building a tiny "nano-rover" to roam the surface of an asteroid, which will be included on a Japanese mission called MUSES-C, scheduled for launch in 2002. He has ideas for true Mini-MAVs, rockets so small they could be ferried around the Martian surface on the backs of rovers--that is, if the payload delivered to orbit were small enough. Then there's the Venus Sample Return concept he and others at JPL have been cooking up. They'd use a balloon to float a golf-ball-size sample up through the thick Venusian atmosphere to an altitude of 40 miles, where they would then fire a rocket from the balloon (just like NOTSNIK did from an airplane) to continue the journey to orbit. Now that, Brian's father might have agreed, is pretty clever--if it works.

Get the latest stories in your inbox every weekday.