low cost innovation in spaceflight the history of the near earth asteroid rendezvous (near) mission

8/8/2019 Low Cost Innovation in Spaceflight the History of the Near Earth Asteroid Rendezvous (NEAR) Mission

1/86

Monographs in Aerospace History No. 36 SP-2005-4536

Low-Cost

Innovationin SpaceflightThe Near EarthAsteroid Rendezvous (NEAR)Shoemaker Mission

Low-Cost

Innovationin SpaceflightThe Near EarthAsteroid Rendezvous (NEAR)Shoemaker Mission


2/86

Cover (from the top): The cover combines a closeup image of Eros, a photograph of the 1996 launch of the Near Earth Asteroid Rendezvous(NEAR) expedition, and a picture of the mission operations center taken during the second year of flight. NEAR operations manager Mark

Holdridge stands behind the flight consoles.

Asteroid and operations center photographs courtesy of Johns Hopkins University/Applied Physics Laboratory.

Launch photograph courtesy of the National Aeronautics and Space Administration. (NASA KSC-96PC-308)


3/86

Howard E. McCurdy

Low-Cost Innovation

in Spaceflight

The Near Earth Asteroid Rendezvous

(NEAR) Shoemaker Mission

The NASA History Series

Monographs in Aerospace History

Number 36

NASA SP-2005-4536

National Aeronautics and Space Administration

Office of External RelationsHistory DivisionWashington, DC

2005


4/86

Library of Congress Cataloging-in-Publication Data

McCurdy, Howard E.

Low cost innovation in spaceflight : the Near-Earth Asteroid Rendezvous (NEAR-Shoemaker

Howard E. McCurdy.

p. cm. (Monographs in aerospace history ; no. 36)

1. Microspacecraft. 2. Eros (Asteroid) 3. AstronauticsUnited StatesCost control.4. Outer spaceExplorationCost control. 5. United States. National Aeronautics and SAdministrationAppropriations and expenditures.


5/86

List of Tables

Acknowledgments

Chapter 1: Challenge

Chapter 2: Origins

Chapter 3: Spacecraft

Chapter 4: Flight

Chapter 5: Legacy

iv

v

1

5

17

35

51

Table of Contents


6/86

Table 2-1 Mass Summary

Table 3-1 Planned Cost of NEAR Mission

Table 3-2 Selected Spacecraft Development Costs

Table 3-3 Actual Mission Costs: NEAR and Mars Pathf

Table 3-4 Cost of NEAR Scientific Instruments

Table 3-5 APL Management Guidelines for Low-Cost M

11

18

18

22

23

31

List of Tables


7/86

In preparing this monograph, the author was assisted by officials at the Johns Hopkin

Applied Physics Laboratory, the Jet Propulsion Laboratory, and NASA Headquarter

to acknowledge the assistance of persons who reviewed the manuscript or otherw

support for its production, especially Stamatios M. Krimigis, Andrew F. Cheng,

Farquhar, Stephen Garber, Roger D. Launius, Jennifer Troxell, Matthew Rochkind, J

Lisa Jirousek, Tatiana Floyd, and Henry Spencer.

Acknowled


8/86


9/86

On a spring day in 1996, at their research centerin the Maryland countryside, representativesfrom the Johns Hopkins University Applied PhysicsLaboratory (APL) presented Administrator DanielS. Goldin of the National Aeronautics and SpaceAdministration (NASA) with a check for $3.6 mil-lion.1 Two and a half years earlier, APL officials hadagreed to develop a spacecraft capable of conduct-

ing an asteroid rendezvous and to do so for slightlymore than $122 million.This was a remarkably lowsum for a spacecraft due to conduct a planetary-class mission. By contrast, the Mars Observerspacecraft launched in 1992 for an orbital ren-dezvous with the red planet had cost $479 millionto develop, while the upcoming Cassini mission to

Chapter 1: Challenge

funding the Near Earth Asteroid(NEAR) mission in late 1993. They the mission to the Applied Physics Lnot-for-profit research and developme

Johns Hopkins University located ocampus between Baltimore, MaWashington, DC. Scientists and enginlaboratorys Space Department ran the

Space Department at APL had come1959 to build navigation satellites fStates Navy. Although increasinglyNASA work, APL scientists and enever managed a major planetary responsibility commonly fell to the JLaboratory (JPL), a NASA Field Cente


10/86

Stamatios Krimigis, head of the Space Department at the Johns Hopkins University Applied Physics Laboratory (rNASA Administrator Daniel Goldin and Maryland Senator Barbara Mikulski a check representing the cost savings acdesign and construction of the NEAR spacecraft. Gary Smith, Director of the Applied Physics Laboratory, stands on

(Courtesy of Johns Hopkins University/Applied Physics Laboratory)

launch station in Florida. The development pro-gram was over, the spacecraft launched and gone.APL officials totaled the money spent on design,testing, flight systems, scientific instruments,prelaunch operations, and project management. Tothe delight of mission advocates, the cost of devel-oping the NEAR spacecraft totaled $3.6 million lessthan the original estimate. Someone suggested that

the project leaders present an oversized, ceremonialcheck to NASA for the money they had saved.Goldin happily accepted it that spring.

Ten months later, on 4 December 1996, anotherDelta 2 model 7925 rocket departed from theCanaveral launch station. It too carried a low-cost

NASAs early efforts to demonstrate of low-cost spaceflight. Both missionsuccessful. The Pathfinder robot landthe summer of 1997. NEAR chasedEros around the solar system for fourorbiting the 21-mile-long object in FeAfter circling Eros for one year, NEAR-Shoemaker spacecraft touched

asteroids surface in February 2001, landing in the history of spaceflight.

Much has been written about theof the Pathfinder mission, including bleaders Brian Muirhead and DonReports on the NEAR spacecraft, the


11/86

agement challenges involved in conducting low-costmissions while daily confronting the possibilityof defeat.

The history that follows is divided into four sec-tions. The first section recounts the origins of theexpedition and the struggle to get the mission fundedand approved. It explains how a small group ofpeople came to believe that an asteroid rendezvouscould be conducted as a low-cost mission, a revolu-tionary proposition at the time. Section two con-centrates on the methods employed to translate the

low-cost philosophy into a robotic spacecraft thatactually worked. Attention is given to the team-building techniques that allowed the people organ-izing the mission to simultaneously restrain cost,meet the launch schedule, and reduce risk. In sec-tion three, the management challenges involved inflying the NEAR spacecraft over the five-year flightregime are described. The difficulties were substan-tial, involving the guidance of a low-cost robotic

spacecraft and the coordination of mission teams atthree different locations. On the first rendezvousattempt with Eros, the little spacecraft missed itstarget, thus requiring another trip around the solarsystem and significant changes in organizationalprotocols. Finally, section four assesses the faster,better, cheaper4 initiative and the NEAR missionscontribution to it.

The scientific returns from the first sustainedexamination of a near-Earth asteroid were impres-sive. Equal to them in importance were the manage-ment lessons learned. For many decades, visionaries

have predicted a future of cheap anflight by which means humans and twould spread into the solar system. Th

sion team undertook one of the pionin that regard, producing a spacecranot only scientific theories concernintion of the solar system, but also manories on the reduction of cost.

In the years that followed, enthufaster, better, cheaper initiative waNASA officials lost four of the five

cheaper missions they attempted toIn 2003, the loss of the Space Shutcaused experts to question the wisdocost-saving techniques to spacecrahumans fly. The expenses associated wmissions rebounded. The faster, bephrase disappeared, replaced by a coOne NASA, a new initiative with nences to low-cost innovation.5

The legacy of the NEAR spacePathfinder mission with which it onnonetheless remain important to the opment of spaceflight. If humanmachines enter space on the scale evarious experts, the movement will nships that each cost billions of dollarwith earlier movements into other r

tious and expensive expeditions wilaffordable technologies. The early eflow-cost spacecraft contain importathe people who work to make this vis


12/86


13/86

To send a spacecraft to an asteroid, where itcould orbit and study the object, required in theminds of most experts a medium-size machine ofthe Mariner-Observer class. The Mariners were aseries of robotic spacecraft, launched between 1962and 1975, that flew to Venus, Mercury, and Mars.Most of the Mariner spacecraft flew by planets; oneorbited its destination. Most of the Mariners were

purchased and produced in pairs. Mariners 8 and 9each had a dry mass (not including fuel) of about2,200 pounds. Given the value of the aerospace dol-lar in the late 20th century, when the NEAR space-craft was produced, each pair of Mariner twins costan average equivalent of $600 million. That sumpaid only for the spacecraft, along with the salaries

Chapter 2: Origins

currency of the late 20th century,approach $1 billion to design, build, la

Observer-class spacecraft coMariner tradition. Although conceiline of spacecraft, only one ObservLaunched in 1992, Mars Observer worbit the red planet and study its clim

ogy. The robotic spacecraft had a mpounds at launch, not counting its fuewas originally conceived as a low-project expenses soared until they than $800 million for the spacecravehicle, and the operations team reamission and conduct experiments.


14/86

stories, presented in a narrative form, in order tomake organizational lessons more memorable andtransmit the cultural beliefs that remind others ofthe hard-learned lessons of experience. As a numberof scholars and practitioners have observed, storiesshape the assumptions that people hold as they goabout their work. In the 1990s, no story had moreinfluence on the people planning NASAs planetaryprogram than the one told about Mars Observer.

The facts of the case are these. Space scientistsin the early 1980s proposed a series of low-costrobotic spacecraft called Planetary Observers thatwould be used to explore the inner solar system.Based on the design of Earth-orbiting satellites, thePlanetary Observers were to provide an inexpensivemeans for conducting studies of Mercury, Venus,and Mars. In 1984, the U.S. Congress authorizedand funded the first project in this series, then calledthe Mars Geoscience Climatology Orbiter, laterrenamed simply Mars Observer. In the beginning,

advocates promised to build the spacecraft for therelatively low cost of $252 million.1 Project scien-tists and engineers designed the spacecraft so that itcould map the surface of Mars, study geologic fea-tures, and provide all-season weather reports. Theylaunched the spacecraft in 1992. Actual spendingon the project totaled $813 million, measured tothe point in the summer of 1993 when the space-craft approached Mars. At that moment, the space-

craft disappeared.

Mars Observer symbolized the folly of the low-cost philosophy. From the date of its conception,the cost of the mission grew. Even so, the missionfailed. People familiar with space missions generallyexplained the results in the following way. Nospacecraft had visited Mars for 17 years, not sincethe arrival of the Viking mission in 1976. The

Soviet Union had attempted two missions but hadfailed, and the last of the Viking orbiters had runout of fuel in 1980. The pent-up desire of scientiststo study Mars led them to attach an ambitiousarray of instruments onto the Observer spacecraft:a high-resolution camera, a thermal emission spec-trometer, a laser altimeter, a gamma-ray spectrome-

The Titan launch vehicle and associtions cost $293 million; this amount,$479 million expense of building the sthe $41 million spent for abbreviated ations, totaled $813 million. Had tcompleted its mission, according to the total cost would have approached

Speaking of the experience, NAtrator Daniel Goldin complained thagrowing complexity and increasingitself, producing less frequent missioexpanding appropriations. It was littobserved, that the space program waport among the Congress, the public, growing old while waiting to fly their

Launching fewer spacecraft meatists want to pile every instrumenonto whatevers going to fly. Thatthe weight, which increases the c

spacecraft and the launcher. Fewcraft also means we cant take anythe ones we launch, so we haveredundancy, which increases wecost, and we cant risk flying newogy, so we dont end up producinedge technology.3

The added complexity of the spac

against the desire of government offstrain the steadily growing proIncreased complexity pushed costs promise to build inexpensive spacecosts down. The resulting equilibrupward and downward cost pressuspacecraft that was a bit too comamount of money allocated to reducThe risks prevailed. Investigators co

difficulties within the plumbing systemdeliver rocket fuel to the flight enginlines to rupture and the spacecraft tcontrol. Within the context of the stoMars Observer in 1993 provided vithose who doubted the wisdom of apcost philosophy to complicated deep


15/86

the advocates of the NEAR mission observed,everybody knows what has happened to that.4

Skepticism ran so deep that some good mis-sions could not get approved. American scientistswanted the U.S. government to send a spacecraftto intercept Halleys Comet, due to return to theinner solar system in 1986. The thought that thiswould require a Mariner-/Observer-class machineof indeterminable cost caused public officials toabandon the mission. The Russians, Japanese, andEuropeans sent spacecraft to visit Halleys Comet,but U.S. officials decided that they could not affordto approve a spacecraft that promised low cost butdelivered cost growth.5

To propose a low-cost comet or asteroid ren-dezvous with a Mariner-/Observer-class spacecraftduring the mid-1980s or early 1990s invited disbe-lief from public officials who approved the funds.To suggest that any such spacecraft be made mas-

sive enough to actually land on a comet or asteroidinvited ridicule. For all its expense and delay, MarsOrbiter could not land on Mars. The only NASAmission to accomplish that objectivethe Vikingprojecthad consumed over $3 billion in inflation-adjusted currency. Public officials simply were notprepared to raid the U.S. treasury for the tax rev-enues apparently necessary to send large, expensiverobots to study chunks of rock or ice.

Could it be done for less? For decades, scien-tists had wanted to visit asteroids and collect sam-ples from comet tails. For much of that time, theyhad been told that the real cost of Mariner-/Observer-type spacecraft was too high. However,the technical feasibility of such a mission was not indoubt. By 1990, NASA spacecraft had visited all ofthe major planets, orbited two, and landed on

Mars. The principal issue in late 19891990,observed the director of the space department atAPL, was the exorbitant development cost.6

Inventing the Discovery

employees specialized in planetary lanbys; as of the late 1980s, they weresend the Observer spacecraft to MGalileo spacecraft to Jupiter. Rather t

JPL in their traditional area of emphasisees turned to the realm of comets and a

Comets and asteroids are the buout of which the planets probably foring more massive bodies, they altered istics of larger spheres and, in the credirected the evolution of life. As pring blocks, asteroids and comets conthe formation of the solar system aportation of substances that give risecontain metals and other resourcesvalue to future civilizations on EarthIf humans maintain their terrestrial cienough, they will need to redirect large comets and asteroids that wostrike the planet with catastrophic res

reasons and more, a special advisory1983 had urged NASA officials to incoid rendezvous in their long-range mission, the Solar System Exploratiohad proposed, might be launched using a spacecraft similar to the employees were aiming toward MaObserver.7 Responding to the expectawould be asked to direct such a miss

at JPL studied the technical requiremcluded that an asteroid rendezvous misible with a spacecraft of the Observe

The Jet Propulsion Laboratory sacre, campus-like setting adjacentGabriel Mountains, just northwest Pasadena, California. JPL was foundSecond World War by a group of p

dents, and associates from the nearInstitute of Technology who perceivedArmy would pay for a research laborto the development and testing of sWhen the space race began, Congrethe operation to the newly creaAeronautics and Space Administrati


16/86

pus twice as large as the main Johns HopkinsUniversity campus 25 miles north of APL. Like JPL,APL was created during the Second World War bya group of university associates seeking to assistwith the war effort. JPL was created to assist theU.S. Army; APL helped the U.S. Navy. JPL workersdeveloped jet-assisted aircraft takeoff; APL scien-tists developed the proximity fuse. JPL is listed as aNASA Field Center; APL is not. Both laboratoriesare associated with major research universities.

Armed with the results of the JPL feasibilitystudy, mission proponents assembled a ScienceWorking Group to advance the concept. To chairthe group, they chose Joseph Veverka, a professorfrom the Cornell University Department ofAstronomy. When he was young, Veverkas familyfled his native Czechoslovakia to avoid the commu-nist takeover. The family settled in northernOntario, where the brilliant night skies inspiredVeverka to study physics and astronomy. Veverkahad been involved in nearly every solar systemexploration project since NASA built Mariner 9,when he was a doctoral student at HarvardUniversity. He was an early advocate of comet andasteroid exploration.

Veverkas group met through 1985 and identi-fied a suitable target, a kilometer-wide body knownas asteroid 3361 Orpheus. The elliptic orbit of

Orpheus causes it to cross Earths path as both bod-ies travel around the Sun. Orpheus is a favored tar-get of space enthusiasts who would like to mine itfor hydrogen, oxygen, and building materials.Veverkas group wanted to visit it. Group membersprepared a list of essential instruments to be flownon a robotic spacecraft and recommended a1994 launch. The working group was dominatedby JPL employees, with whom Veverka had worked

on the Mariner, Viking, and Voyager projects.Significantly, no representatives from APL appearedamong the outside members. The working groupgave the project a name: Near Earth AsteroidRendezvous (NEAR).8

Veverka reappeared in the summer of 1989 at a

chaired the plenary session in which of scientists made a radical proposinitiative in low-cost spacecraft.

Up to that point, participants hadany new initiative for a low-cost planwould involve an Observer-class s1989, however, Mars Observer was wget and behind schedule. The desireplanetary spacecraft that could be preperiods of time clashed with the lessoObserver. As part of the workshop, aMission Program Group was asked rationale for a new mission line that wcost and provide a mechanism wherebentific questions could be addressed short time. The initiative, Veverka rgreeted with widespread skepticism.

Stamatios Krimigis did not sharsense of disbelief. A 21-year veteKrimigis was the Chief Scientist in thDepartment. Krimigis was born aChios, Greece, immigrating to the Unstudy physics at the University of Minearning his bachelors degree, he enUniversity of Iowa, where he studinotable James Van Allen, famous forscientific instruments on the first U.Sing satellite that discovered the radiati

ing his name. Krimigis wrote his disolar protons and received his doctoLike Veverka, who would become Astronomy Department at CornelKrimigis eventually would become Space Department at APL. Both were gent individuals accustomed to tprocess of drawing consensus from and often irascible colleagues.

Krimigis suggested that the SmProgram Group not use Mars Observof its quest for low-cost solar systemRather, he pointed to a long standingram called Explorer that had servedthe space physics community and had


17/86

pared by Wernher von Braun and his German rocketteam, and an experiment designed by Krimigissmentor at the University of Iowa.11 The Explorer 1mission was developed in an incredibly short periodof time. The Von Braun-JPL-Van Allen team organ-ized the mission in just 84 days before its 31

January 1958 launch.

Between 1958 and 1981, civilian space officialsorganized some 60 Explorer-class missions. Mostwere designed to study phenomena like the solarwind, cosmic rays, and Earths magnetosphere. Theexact number of missions is hard to identifybecause Explorer-class missions often bear nameslike Vanguard, Ariel, and Boreas, and some satel-lites named Explorer (such as the CosmicBackground Explorer) are too large to fit the cate-gory. Strictly speaking, Explorer-class satellites aresmall and uncomplicated. Explorer 1 had a mass ofjust 22 pounds, while later versions (like Explorer58) totaled about 250 pounds. With the advent ofthe Space Shuttle in 1981, rated to carry up to65,000 pounds, the number of small Explorersdeclined in favor of larger, multipurpose (and hencemore expensive) platforms.

At the time of the 1989 workshop, Krimigiswas working on a concept called the AdvancedComposition Explorer (ACE). This spacecraft,eventually launched in 1997, was designed to fly

one million miles from Earth and study the solarwind. The spacecraft would have a mass of 1,300pounds, small for such a distant traveler. Krimigispredicted that the ACE spacecraft would be builtfor less than $100 million. It would be designed andassembled using a simple management model, headded, in which a single Principal Investigator over-saw the development of the spacecraft and each ofthe lead investigators provided their own scientific

instruments.

Geoffrey Briggs, head of NASAs Solar SystemExploration Division, listened intently. Heapproached Krimigis after the presentation andasked if APL would be willing to turn the conceptof an Explorer-type spacecraft into a proposal for a

The oversight group met in 1989 amembers did not recommend a specifithey did propose a name for the overlow-cost initiatives: they called it tProgram.

In 1990, Wesley Huntress replahead of the Solar System ExploratioNASA Headquarters. Huntress had spyears as a research scientist at JPL, wof the general preference for large psions, he had developed a commitmention of cost. Among his scientific interspecialized in asteroids and comets. Hganized the Science Working Group, rthey select a mission and test the fealow-cost approach. As part of the reHuntress selected a new chair forWorking Group. He asked Joseph Vethe group; Veverka agreed.

Selecting aReducing cost while improving p

difficult, but not impossible. As the electronics and computer industries goal requires someone to design macsmaller and easier to operate. The re

vacuum tubes with transistors, allowed industrialists to design radiobetter and cost less. One elementary moperational complexity of spacecraft cby examining the number of mOn Mars Observer, for example, thmoved. So did the high-gain antennacomplexity of Mars Observer resulted with a mass of 2,240 pounds at laun

propellant. A 2,000-pound mass wasof the Mariner-/Observer-class spacec

Explorer-class spacecraft might hathat mass, but they were not designedmissions. If clever engineers could cutspacecraft designed for Mariner-/O


18/86

the Discovery Program emerged, additional con-straints appeared. The total cost of operating thespacecraft, once launched, could not exceed $35million, and the spacecraft had to be small enoughto fit on a rocket no larger than a Delta II, which,in the 1990s, cost about $55 million per launch.12

At the Maryland laboratory, Stamatios Krimigisassembled his staff and appointed a small group tostudy the NEAR concept. The group contained old-timers and youngsters, people from both inside andoutside the Lab. Krimigis appointed himself as thelead scientist. He selected Robert Gold to study theinstrument package. Gold, a physicist with 10 yearsof experience in the field of remote sensing, hadjoined APL in 1977 and had served as the leadinvestigator for two of the instruments on the ACEspacecraft. Krimigis asked Edward Reynolds toserve as systems engineer. A relative newcomer,Reynolds had joined the Lab in 1985, performingsystems integration work on complex nuclear elec-tric propulsion and strategic defense radar pro-grams. Andrew Cheng, a young physicist who hadcompleted his doctorate just three years earlier,agreed to serve as the chief scientist. Cheng hadserved on a number of NASA advisory committeesbut was not a member of the APL staff at that time.To study the mission design, Krimigis recruitedRobert Farquhar, a brilliant aeronautical engineerwith 30 years of experience in the field. Farquharhad spent most of his career at NASA and wasin the process of moving to APL. As the study man-ager, Krimigis selected Thomas Coughlin, a soft-spoken mechanical engineer. Coughlin had joinedAPL in 1982 and participated in several low-cost,fast-track Strategic Defense Initiative programs,including the renowned Delta 181 project on whichhe had served as the program manager.

To make the mission worthwhile, the groupinsisted that the spacecraft carry at least two instru-ments, a multispectral imager and a gamma-rayspectrometer. A multispectral imager is a sophisti-cated camera, while a gamma ray-spectrometersenses the composition of objects under study. Wehoped we could eventually accommodate more

be sophisticated enough to deliver thon what would be a multiyear voyage ing the $150-million development goa

On the other side of the country, a similar group studied the problem

JPL. Unlike the people at APL, worked at a facility with no extensivlow-cost exploration. The cost of Jprojects typically approached or exclion; the upcoming Cassini mission to top $3.3 billion. The two teams appeaVeverka working group in May 1991Pasadena, California, home ground two teams, observed Krimigis, had tlingly different conclusions.14

It is improbable, JPL team membthat a near-Earth object rendezvous ried out for less than $150 million, evple, no-new-technology machine.15 Timprobability, the JPL team made an igestion. By building and flying threteam members observed, the cost assigned to the first spacecraft could bmillion. The funds assigned to the secspacecraft would be slightly less: $14$137 million, respectively. Total coexplained, would be $436 million.16

This was an ingenious approachproduce a Mariner-/Observer-class spon the cost accounting sheets, lookething less. The JPL team sought to taof the cost benefits inherent in a prwhile reaping the accounting benefitsassigning project expenses to three interdependent missions. The incremof producing a second and third spac

to the first is typically small, an advateam sought to exploit. On the VikiMars some 15 years earlier, the incremnot producing a third lander amountemillion out of a total spacecraft deveget of $357 million. The JPL team cone flight for $150 million, but that


19/86

Based on the positions of the ESun, and the target asteroid (Athat time), we convinced ourseNEAR could be configured withigh gain antenna, fixed solar paninstruments with a common orthogonal to the antenna axis, anengine thrusting opposite to the iboresight. With this simple spaccould accomplish all the science of the mission.18

The APL team agreed that theysuch a spacecraft with a dry mass of 430 kilograms (946 pounds). Their ations totaled 363 kilograms, plus areserve. On top of that, they added a allowance for fuel. The groups abilplish the mission with such a spacecupon a complex relationship among mass, the lifting power of a Delta 2 lavarious launch window opportunitiestrajectories, and the energy required foin velocity necessary to maneuver theits destination. The group was reasothat their design could accommoda

Officials from the Applied Physics Laboratory prepared thissketch in early 1991 to demonstrate how a simple spacecraftmight carry out an asteroid rendezvous at a very low cost.


three missions could easily reach $50 million ormore. Although the JPL team gave no total missionestimate, one can surmise that the overall costwould top $600 million. Such an amountapproached the $813-million sum that JPL officialswere spending at that time on Mars Observer,which would be launched the following year.

Members of the APL team rose to explain theirapproach. Tom Coughlin directed the presentation.

The groups proposal, to say the least, was quitedifferent. The great pleasure in life is doing whatpeople say you cannot do,17 team member RobertFarquhar later observed. In January, the APL teamhad prepared a sketch outlining a concept for aremarkably simple spacecraft. The drawing showeda cylindrical spacecraft shaped like a large tuna can.

Mass Summary

Instruments

Propulsion

Power

RF

Attitude

C & DH

Thermal

Harness

Structure

Dry Mass

18.5% Contingenc

Fuel

Table 2-1


20/86

geted missions . . . based on Delta II performance,launch window considerations, and missionrequirements such as the maximum velocity changesthat the spacecraft engines could perform.19

An accompanying chart laid out the spacecraftdevelopment schedule. If the project were approvedsoon, in 1991, detailed design and fabrication couldbe completed in four years. The spacecraft wouldbe ready for launch in May 1997. The schedule, oneAPL study group member admitted, was extre-mely aggressive.20 Finally, the APL group agreedthat they could design and construct the spacecraftfor less than the $150 million target price set by theVeverka group. In fact, they offered to build the space-craft for $40 million lessa mere $110 million.21

Andrew Cheng remembered a general reactionof not so much skepticism as amusement.22 Agroup of people with little experience in conductingplanetary missions was offering to build a simplespacecraft and fly it millions of miles to an object oflimited size. Moreover, members of the group saidthat they could build and test the spacecraft in lessthan four years and for less than $150 million.The reviews were mixed, said one of the peopleattending the meeting. Many expressed disbeliefthat such a mission could be done.23 Although offi-cials in the APL Space Department had a goodrecord of delivering low-cost projects on time,space projects within NASA as a whole were expe-riencing substantial cost overrunsan average of77 percent over original estimates, according to astudy completed a few years later.24 The peoplereviewing the two proposals did not want toapprove a low-cost project and then watch its costgrow once work began.

Joseph Veverka, the Cornell astronomer head-ing the reviewing board, was intrigued by theAPL proposal. He had ties to JPL but favored thelow-cost approach. Cheng recognized that if wecould succeed in this approach, we had a winningconcept.25

The leadership at JPL viewed the potential for

Mercury.26 JPL executives had coNASAs Ames Research Center California and prevailed; they did notcompetition from APL.

Edward Stone, the new DirectoPropulsion Laboratory, told his peoptheir concept. One month later, theWith a small team of believers, reported, a $150 [million] discoveachievable at JPL. They were ready on an asteroid rendezvous mission We have the experience, capability manage this Discovery project, the r

The study teams new position wcant than the name of the personpresentation. JPL executives assignedSpear to present the new findings. Spchallenge the traditional JPL culture, twith the Cassini mission to Saturn, bCenter toward smaller and less expenAt the time, he and other dissidents for a series of very low-cost spacecraestablish a network of environmentastations across the planet Mars. Thecalled the Mars Environmental SurvIts advocates envisioned a network craft, the equivalent of meteorologicalat various locations. To land such a laspacecraft on such difficult terrain reqtionary technologies and severe cost c

On 17 June 1991, Spear issued hito the Veverka group on behalf of amission. Veverka reported to Weswho, like Veverka, had spent much working with people at JPL. The astwas based on the proposal of a Working Group that had been domiemployees. JPL had more experience ietary missions, but APL had a betterwith low-cost missions.


21/86

solar system. According to the guidelines, each mis-sion in the Discovery Program would cost no morethan $150 million, perhaps less, for spacecraftdesign and development. For the first mission in theprogram, Veverkas group recommended an aster-oid rendezvous. Although the group members leftthe selection decision to NASA, their recommendedguidelines contained the essential elements of theAPL plan.

The first mission of the Discovery Programshould be a rendezvous with a Near-Earth aster-oid, the group began. The most suitable destina-tion was asteroid 1943 Anteros, the target forwhich the APL team had been planning. TheVeverka group recommended that the spacecraftcarry three instruments, including an imagingdevice, and be small enough to launch on a low-costDelta 7925 rocket. The latter imposed a severe con-straint. To assess whether a spacecraft can completeits mission, planners calculate all of the changes invelocity (called V) necessary for the vehicle toleave its low-Earth parking orbit, move throughspace, rendezvous with its destination, and com-plete its mission. Generally speaking, a higher Vrequires a larger, more complex (and hence moreexpensive) mission. The constraints imposed by theVeverka group limited the mission to velocitychanges totaling no more than 5.5 kilometers persecond (12,296 miles per hour).

Including the boost necessary to move thespacecraft out of low-Earth orbit, mission plannerscould expect to make more than 50 velocitychanges to reach and study a near-Earth asteroid.Most proposals for asteroid rendezvous missionscalled for total velocity changes in excess of 6 kilo-meters per second. Reducing that sum required asmall spacecraft and a clever trajectory, exactly theapproach that the APL team said they could carryout.28 The APL team calculated that they couldreach Anteros with a total V not exceeding 5.35kilometers per second.29

Simultaneously, in a show of support for theconcept, members of the Senate subcommittee

Maryland-based laboratory. The comin which this language appeared urgecials to emphasize projects which coplished by the academic or research ca way of saying that institutions like Agiven a chance to play.30

On 1 April 1992, NASA acqAdministrator, Daniel S. Goldin. Aexecutive, Goldin had been an advolow-cost spacecraft. The White House selected him wanted someone who NASAs partiality for large, expenGoldin had participated in the Briinitiative, a plan to mass-produce inexpsensors that could detect a missile attaproof precision. Once Goldin arrivhe found that the realities of fiscal retreported his position. The Agency simafford expensive new initiatives. Goldhis own words, to re-invent NASA:

Lets see how many [spacecraftbuild that weigh hundreds, not thof pounds; that use cutting-edgeogy, not 10-year-old technology it safe; that cost tens and hundrelions, not billions; and take moyears, not decades, to build andtheir destination.31

Shortly after arriving at his newdiscovered the fledgling Discovery was very impressed. He called it thkept secret and became one of itscates.32 Pushing from the top, he chalemployees to cut mission costs. Build craft, he urged Agency employees. Mtrol as important as scientific dischances, he said. A project thats 2successful, he observed. Its prooplaying it too safe.33

Shortly after Goldin assumed leacivil space agency, officials in NASSpace Science and Applications u


22/86

cussed by the scientific community. Today they arethe centerpiece of NASAs new programs for the1990s. The report announced the Agencys intentto promote a new program called Discovery that

would produce a series of small, cost-effectivespacecraft modeled on NASAs existing Explorerand Earth Probe programs, just as the APL teamhad proposed.34

To the shock of the APL team, however, thereport did not recommend an asteroid rendezvousmission as the first Discovery project. Rather, thefirst flight award went to Tony Spear and the JPLteam promoting MESUR. Spears group promisedto launch the first spacecraft in the MESUR seriesby the end of 1996, thus providing an earlier test ofthe low-cost concept than would be possible underthe APL proposal for a 1997 launch to the asteroidAnteros. While government officials never approvedthe whole MESUR concept, they did authorize thefirst spacecraft. It was called Mars Pathfinder. TonySpear served as the project manager for the Pathfindermission as the spacecraft was designed and built.

What about the NEAR mission? It was a sec-ond concept under study, according to the authorsof the small planetary mission plan.35 Under theterms suggested by Veverkas Science WorkingGroup, the NEAR spacecraft would be launched inMay 1997 as the Pathfinder spacecraft approachedMars. Within a few months, that launch date hadslipped to 1998. The later launch date necessitateda new targetthe asteroid known as 4660Nereus.36 Compared to Anteros, Nereus was a lessinteresting target. It was comparatively puny, just 1kilometer in diameter. Members of the NEAR teamworried about the change. Although Nereus satis-fied all the requirements for a Discovery-classmission, three of them wrote, some scientistswere concerned that its small size could restrict thequantity and diversity of the science return. Putanother way, the three observed, the results would besomewhat boring.37 As the Pathfinder expeditiongained momentum, the asteroid rendezvous projectslipped away.

ball, said one of his colleagues, anfirst to admit it.38

As a civil service employee at NA

Space Flight Center in Maryland, Fparticipated in the InternationaExplorer-3 (ISEE-3) mission as its director. This otherwise inconspicumade spacecraft history when Farquhtrajectory that allowed it to beat thRussians, and Japanese to the first enccomet. Following Farquhars plan, flignudged the ISEE-3 spacecraft out of tion point where it was conducting itallowed it to fall toward the MooClarke, the famous science fiction wrresulting maneuver one of the mosfeats of astrodynamics ever attempteda path that looked like the curled childs birthday present, the spacecrgravitational boost each time it flew This maneuver was repeated no ltimes; on the last flyby, the probe sa hundred kilometers above the mslightest navigational error could havtrous,39 wrote Clarke. The lunarconsisting of incredible twists and wmonths to complete and eventuallyISEE-3 spacecraft in the right directioa nearly two-year journey to the comZinner.40

Farquhars ultimate boss at the Malation, Goddard Center Director Noel Hacterized him as a very bright, very innSaid Hinners: I would like to havehim around herenot 100, we cothat.41 Farquhar is something of a outsider prone to irritating people lethan he. Hes an idea man, obresearch scientist Donald Yeomans. Hideas, 99 of which may not be reasonlast one is a beaut. And he does this r

The launch window for any dezvous was dictated by the position


23/86

fiscal year ending in 1994. In the wiNASA executives asked Congress for both the NEAR and the Pathfinder pHouse of Representatives, where vo

tioned by population, lawmakers funds for the California-based Pathfbut refused to start funding the Masteroid rendezvous.

In the Senate, where each state representation, Maryland is as big Though diminutive in stature, Barbwas a large presence on the Senate ACommittee. With a small amount of hexecutives, Mikulski placed $62 milliappropriation bill for the purpose oNear Earth Asteroid Rendezvous pconference committee negotiations tMikulski prevailed.

On the grounds of the Maryland 5 December 1993, Senator BarbaStamatios Krimigis, Wesley HuntreDirector Gary Smith announced thabegin work on a Near Earth AsteroidIt would be the first launch in Discovery Program of faster, betspacecraft, the first mission to renan asteroid, the first U.S. planetaryconducted by a non-NASA space cesignificant economic milestone for The mission would be launched in Fon a Delta 2 rocket and would arrisubsequent NASA announcement expDecember 1998.

When the celebration died dowassembled members of the NEAR spconfronted a sobering reality. They haflight-ready, fully workable spaceFlorida launch site by December 1months away. This was an utimetable for developing a planetary mteam leader Thomas Coughlina sevconcept that a planetary expedition together cheaper, better, and faster, to

second), a conclusion endorsed by the Veverkagroup. In the same paper, Farquhar noted that aspacecraft launched in 2003 with the assist of anEarth swingby could reach the asteroid Eros by

2005 while flying by Enckes comet on the way.Eros was especially interesting, Farquhar noted,because it was the largest near-Earth asteroid.43

Nereus is a mere kilometer wide, while Eros has amean diameter 20 times that large.

Farquhar reexamined his calculations. A space-craft could reach Eros at the end of what he calleda small-body grand tour of two comets and twoasteroids.44 This approach allowed an earlierlaunch date in 2000, but the spacecraft would notreach Eros until 2012. Farquhar performed thecalculations again and again. Eros, he noticed, waspositioned in such a way that a launch in early1996, when coupled with an Earth flyby, couldplace a spacecraft at the asteroid in about threeyears. It was a long flight, but the energy require-ments worked. It could be accomplished with veloc-ity changes that totaled just 5.53 kilometers persecond, substantially less than those required byother Eros opportunities. Significantly, theFebruary 1996 launch window allowed the APLteam to beat the Pathfinder group to the launch padby 10 months. The Pathfinder spacecraft wouldarrive at Mars long before NEAR reached Eros, butFarquhar also had a solution for that problem. Byadjusting the NEAR flight path, the APL missionteam could achieve an asteroid flyby on 27 June1997, one week before Pathfinder reached Mars.The arrival date at Eros was unchanged, but the

June flyby would give scientists a first quick look ata large, 61-kilometer-wide, carbon-rich asteroid.

With this mission plan, the NEAR projectcould be first again. Farquhar took the plan toStamatios Krimigis, head of the APL SpaceDepartment since early 1991, who called it anexciting opportunity.45 Together, they transmittedthe proposal to NASA officials in Washington, DC.

To make the mission work, Congress had toappropriate funds to start the project immediately,


24/86


25/86

The Applied Physics Laboratory team wonapproval to begin work on the Near EarthAsteroid Rendezvous mission in late 1993. Gainingapproval for the project had been difficult; creatingthe spacecraft would be harder. Members of theteam had to design, build, and test a robotic space-craft that actually workedand do so within thecost and schedule constraints they had set for them-selves. The mid-February 1996 launch window satjust 26 months away, an impossibly short period oftime for the preparation of a planetary-type expedi-tion. The $214.5 million that NASA officialsplanned to spend on the entire mission (spacecraft,launch vehicle, and flight operations) was consider-ably less than teams normally devoted to ventures

Chapter 3: Spacecraft

As a point of contrast, the missioing on the Mars Observer spacecraf1992, had spent seven years preparincraft after their official funding staObserver team had devoted $479 micraft development. Even counting thethe May 1991 presentation to thScience Working Group, the NEAR tethemselves significantly less time. AdNEAR team had promised to finish tof the projectconstruction of the spa paltry $122 million, a considerableto the low-cost philosophy. (See table

Nor were they building an ordin


26/86

When it arrived at its destination, the spacecrafthad to enter an orbit around an object possessinglittle mass. The spacecraft team would need to flythe tiny spacecraft around the asteroid in orbits ofever-decreasing altitude, skimming over the surfacewithout crashing into it.

During the 1990s, NASA officials commis-sioned a substantial number of planetary and solarsystem missions. Based on the funds devoted to thedevelopment of the spacecraft, NEAR sat near thebottom of that list. (See table 3-2.) Among majorsolar system initiatives, including other faster, bet-ter, cheaper missions, only Mars Climate Orbiter

sat lower on the scale of money devocraft development, and it had failed.

Had the NEAR team taken on too

ising to build and test a spacecraft tootoo little money? Team leaders were chad not. They did not believe that thon a task that would prove too fast They planned to draw upon more thaexperience in managing low-cost phave a hard time convincing people thnew to us, said one APL project man

Across the continent in CalifornPathfinder group faced a substantichallenge from the one confrontinteam. The NEAR team had extensiwith low-cost spacecraft but little faplanetary-type missions. The Pathworked at an installation with tradition of planetary exploration but commitment to inexpensive spacecr

ment costs for JPL planetary spacecrran toward $1 billion. Many JPL collthe work of Tony Spear and his smteam with disbelief. The notion that a

JPL engineers and scientists could devcraft in less than three years for underand land it on the surface of anothcontrary to the history and culture oMost people thought it could not be d

A different attitude prevailed at Aexcessive pessimism persisted at JPL what might be characterized as excesbubbled forth in Maryland. The ApLab had a long tradition of buildinspacecraft in short periods of timeDepartment at APL had come into exipurpose of building inexpensive navlites for the U.S. Navy. When the missile submarines were built in theinertial guidance navigation systems teTo locate their position at sea accurasary requirement for accurately firimissile), submarine captains had to

Source: Applied Physics Laboratory, NEAR Mission Plan vs.Actuals ($M), undated, folder 8722, NASA HeadquartersHistorical Reference Collection, Washington, DC.

Table 3-1

Planned Cost of NEAR Mission(in millions of real-year dollars)

Mission Elements Cost

Spacecraft 122.1

Mission operations after launch 46.2

Headquarters support 2.7

Launch vehicle 43.5

Total mission cost (planned) 214.5

Table 3-2

Selected Spacecraft Development

Costs (in millions of real-year dollars)

Spacecraft (Launch Date) Cost (Actual)

Cassini (1997) 1,422

Mars Observer (1992) 479Pathfinder (1996) 200

Mars Global Surveyor (1996) 131

Stardust (1999) 127

Near Earth Asteroid Rendezvous (1996) 113

Mars Climate Orbiter (1998) 80 (est )


27/86

ing to create a credible deterrent to nuclear war.The Applied Physics Laboratory developed whatbecame known as the Transit navigation satellitesystem.

The U.S. Navy insisted that Transit satellites beso small that military officers could launch them onScout launch vehicles. Scout rockets were tiny rock-ets, hardly more than guided missiles, capable ofpushing objects weighing little more than 400pounds into space. Although they were mobile andeasy to launch, their use created a severe constrainton the people designing the satellites. APL engineers

responded by producing satellites that weighed but119.5 pounds yet had an overall operational relia-bility rate of 99.9 percent.

Buoyed by their success with the Transit satellitesystem, APL employees offered to prepare and test alow-cost interceptor for the Strategic DefenseInitiative (SDI), popularly known during the 1980sas Star Wars. In a room full of defense contractors

discussing half-billion-dollar systems and five-yearpreparation times, APL representatives proposed atest that would require little more than $150 millionand a year of work. According to one of the peopleat the 1985 meeting where the concept was originallyproposed, the APL representatives were warned thatyoure going to embarrass the United States and thePresident.3 The successful 1986 test of the APLDelta 180 interceptor became the foundation for thebelief among top administration officials that spacesystems did not need to be expensive to work.

The prime responsibility for developing theNEAR spacecraft fell to Thomas B. Coughlin. Hewas one of the most experienced project managersat APL, and his participation in several fast-trackSDI projects had prepared him for the technicalchallenges of producing the NEAR spacecraft injust two years. Coughlin had a clear vision of themanagement philosophy necessary to produce alow-cost, short-schedule spacecraft, but in his typi-cal fashion, he would let other people claim creditfor the new approach.

ments), and Andrew G. Santo as the stem engineer. In all, about 30 people pmembers of the core group responsibing the spacecraft and preparing for f

Aerospace engineers commonlyphilosophy that cost, schedule, and factors to be traded against each othefast-schedule project, according to thshould incur additional risk. To reduagers of fast-track projects typically money. Traditionally, spacecraft mwhat is characterized as a pick tw

toward cost, schedule, and reliabilitheir sponsors to pick two and pay with the third.

The philosophy underlying the cheaper approach challenged this NEAR and Pathfinder teams soughspacecraft that were simultaneously cable and to produce them rapidly. Th

sought to achieve this goal through shigh-tech instrumentation, reduced calculated risk-taking, and team-bament techniques.

Designing for SiThe immediate problem confron

was spacecraft design. Participants oproject had to design a spacecraft thenough to accomplish its mission enough to meet its cost goals. The twoment schedule significantly limited thefor systems integration and testing. spacecraft, individual components liksion or telecommunication subsystemsfairly wellso long as they are tested sWhen linked together, however, they hto fail. Engineers use systems integratito overcome these interactive flawsspacecraft schedule did not leave msystems integration and testing. To red


28/86

timeters) tall. On one end of the box, the teamhard-mounted a 1.5-meter high-gain antenna, theprimary means for communicating with Earth, andfour solar arrays, each 6 feet long by 4 feet wide.

The solar arrays folded down for launch, thenextended into a fixed position once in space.Following their deployment after launch, the solararrays did not move.

On the opposite end of the spacecraft, teammembers hard-mounted four scientific instruments.A fifth instrumentthe magnetometersat at theopposite end, above the high-gain antenna. The

propulsion subsystem consisted of 11 smallthrusters, 1 large bipropellant engine, and accom-panying fuel tanks. The propellant subsystem con-sumed nearly all of the volume inside the spacecraftand essentially dictated its size. The main engineprotruded in a visually awkward fashion throughone of the panels on the octagonal spacecrafts side.

Fixing the high-gain antenna and solar arrays

to the body of the spacecraft vastly simplified itsdesign. On other planetary explorers, these mecha-nisms swiveled so as to allow antennas to point atEarth and panels to point toward the Sun. Duringmost of its multiyear journey through the solar sys-tem, the NEAR spacecraft was designed to travelwith its solar arrays aligned toward the Sun. In thatconfiguration, it communicated with Earth througha fanbeam antenna, an ingeniously powerful device

that resembled a large domino. When the spacecraftneeded more rapid communication, it turned andpointed its dish-shaped high-gain antenna towardEarth. For those periods of the mission, the solarpanels diverted from the Sun at an angle notexceeding 40 degrees. To compensate for the lesserlight falling on the arrays during those periods, themission team changed the material out of which thearrays were constructed and increased their size byan appropriate amount.

The instruments mounted to the opposite enddid not point forward, along the spacecrafts usualpath, but were fixed to look out from the side, likepassengers sitting in an airliner. As if to confuse the

The two-year development scheforemost design driver for the spacmechanisms, team members noted.5

had to be designed and manufactured

ably short period of time. Team leadebegin assembling the whole spa18 months after funding began. Mechof the whole spacecraft was set for the(The team planned to conduct adat Kennedy Space Center before launcunexpected interactions once thwas assembled, designers minimized tcommon hardware. This design,

team members reported, allows paradevelopment, test, and integration compressed spacecraft development anschedule can be met.6 The propulsiwas so independent that two of the ening on the project characterized it mrate spacecraft than a subsystem.7

The simplified design reduced c

and helped prevent unexpected intelems as the launch date approached. Ia spacecraft that was a bit hard to hanvelocity adjustment thruster stuck outthe spacecraft, perpendicular to theback to Earth and the Sun. Misaligthruster during flight maneuvers causcraft to torque. To counteract anyspacecraft would have to burn p

Precise tracing and control of the cenbecame mission critical, lead enginee

The NEAR team spent $79 milliconstructing, and testing its spacecrascientific instruments. In California, tteam produced a lander for just $135table 3-3.) These were small sums bstandards. Most impressive was the sment time. An analysis of NASA spsuggested that planetary missions reqage of eight years of preparation; review of spacecraft difficulties idencient development time as a primary sion failure.9 The NEAR spacecraft te


29/86


30/86

Saving Through Miniaturization

The NEAR spacecraft team saved money byreducing the size of spacecraft components, partic-ularly the instruments placed on board. So did thePathfinder team. Much of the philosophy underly-ing the faster, better, cheaper initiative reliedupon the revolution in solid state electronics.

Throughout the latter part of the 20th century,industrialists used technological advances achievedthrough miniaturization to simultaneously improvethe cost, production schedule, and reliability ofcomputers, calculators, cameras, telephones,radios, and other electronic equipment. To NASAAdministrator Daniel Goldin, microtechnologieswere the sword that would slice through theGordian knot of big, expensive spacecraft that take

forever to finish.10

The camera or imaging system on the NEARspacecraft provides a good example of how suchinnovation can occur. Compared to the cameras onthe Viking spacecraft that landed on Mars some 20

gle pixel at a time. A nodding mirrortion with a set of lenses concentratelight on an array of diodes capable offerent focal lengths and colors, thenprocess for the next dot. The array light beam, whereupon the mechanithe signal, converted it into digitalstored it for transmission back to Ear

era took pictures by scanning a verpixels tall, pixel by pixel, then rotatinanother line. A color panorama took autes to prepare. Scientists anxious to Mars joked that any creature scamperlandscape would appear as a small dissingle column of dots.

Viking project managers hoped

camera for $6.2 million while develotechnology. Technical problems causpend $27.3 million by the time thflew. That is the equivalent of approxmillion in the purchasing value of the1990s, more than the NEAR team pla

Actual Mission Costs: NEAR and Mars Pathfinder(in millions of real-year dollars)

Mission Elements NEAR PathfinderSpacecraft

Flight systems 79.1 135.3

Science and instruments 15.4 13.7

Project management 2.4 7.1

Integration and prelaunch operations 9.8 10.0

Other 6.8 4.6

Subtotal 113.5 170.7

Microrover 25.0

Mission operations after launch 60.8 14.6

Headquarters support 2.7 4.8

Launch vehicle 43.5 50.3

Total 220.5 265.4

Table 3-3

Sources:

Applied Physics LEarth Asteroid RenPhase C/DDeveundated, folder Headquarters HisCollection, Washin

Applied Physics La

Mission Plan vs.undated, folder Headquarters HisCollection, Washin

NASA, Mars PaCost and Fundundated, folder 162quarters HistorCollection, Washin


31/86

At its full field of view, the CCD captured an image537 pixels wide by 244 pixels high. By the mid-1990s, this was not a new technology. EdwardHawkins, who led the APL team building the

device, based much of its design on the wide-fieldimaging system prepared for the flight of theMidcourse Space Experiment (MSX) spacecraft.Hawkins was a relative newcomer at the AppliedPhysics Laboratory, having joined the staff justseven years earlier, and as the NEAR project began,he was still working on his doctorate in physics at

Johns Hopkins University. He had helped developthe MSX imaging system, which APL had com-

pleted for the Ballistic Missile DefenseOrganization. The MSX spacecraft used a multi-spectral imager for the purpose of distinguishingpotential ballistic missiles from the natural spacebackground. It was launched in April 1996, shortlyafter NEAR, but work had begun in 1988, givingthe MSX imager team far more time to perfect tech-nical details. Said Hawkins:

Fundamental to the success of NEAR wasthe development of instruments based onproven flight designs. By adapting existingdesigns and concentrating on and improv-ing the important features, we reduced thedevelopment cost and risk.11

Hawkins anticipated the possibility that thefocal plane detector (FPD), which contained the

CCD and the timing control mechanism, mightinteract with the imaging systems data-processingunit (DPU) in unexpected ways. If that happened,fixing the problem would use up time needed to testand calibrate the flight instrument, which coulddelay the assembly and testing of the whole NEARspacecraft. To prevent that possibility, Hawkins andhis instrument team used an interface between theFPD and DPU that was very similar to the one used

on the Midcourse Space Experiment spacecraft.The use of existing, reliable ground support equip-ment to test the flight hardware was probably thesingle most important factor that enabled theinstrument team to meet the tight schedule of theNEAR program, Hawkins reported.12

The Pathfinder team used the samnology to construct a camera in the focylinder about 4 inches in diameter. Tsystem was constructed at the Univers

Lunar and Planetary Laboratory untion of astronomer Peter Smith. Tteam spent $7.4 million on their cspent much of the money improving calibration of the camera, which Pathfinder operations team to prodtures with accurate colors and threimages.

In all, the NEAR spacecraft teamillion on scientific instruments. (Tteam spent $14 million.) The most expment on the NEAR spacecraft wagamma-ray spectrometer, the primadetermining the surface compositiongamma-ray detector consisted of acylinder, while the x-ray detector toothree tube-shaped devices, sitting side

about 7 inches in length. (See table 3-

The scientific instruments on the craft were small and, by traditional spdards, relatively inexpensive. From a were hardly visible. In essence, the sprocket engine (with accompanying fu

Table 3-4

Cost of NEAR Scientific

Instruments(in millions of real-year dollars)

Instruments Cost (E

Gamma-ray spectrometerSpectrograph

Imager

Magnetometer

Instrument data-processing units


32/86

Technicians mounted scientific instruments on the section of the spacecraft that would point away from Earth anpanels and the high-gain antenna were fixed to the opposite end.


solar arrays) that carried scientific appendagesattached to its outer frame. The ability to miniatur-ize scientific instruments based on existing microelec-tronic technologies was a major factor allowing theNEAR team to achieve its cost and schedule goals.

Reducing Launch CostsFor many years, exploration advocates have

touted the benefits of cheaper access to space.Imagine launch vehicles that could carry payloads

In spite of such hopes and promisorbital access remained stuck througha range that varied from about $4,5per pound. (That provided access to Lthe other planets and asteroids costmore.) To achieve really cheap access sion managers were left with only o

reducing launch costs: they had to redsize and mass, thereby flying on smexpensive launch vehicles.

As a condition for approving theect, NASA officials confined the AP


33/86

for keeping the spacecraft small. The NEAR teamspent nearly $25,000 per pound to dispatch itsspacecraft and the accompanying propellanta large sum, but not unusually high for machinery

sent on solar system expeditions. Overall launchcosts, however, barely broke $43 million, sincethe mass of the fully fueled spacecraft as it sat onthe launch pad totaled only 1,775 pounds (805kilograms), including 717 pounds (325 kilograms)of fuel.

The requirement for a small launch vehicleforced the NEAR team to miniaturize the space-

craft, itself a means of cost control. Spacecraft masswas always a concern, said the NEAR spacecraftengineers. They worried about spacecraft mass untilthe official weighing operation a few days beforelaunch.14 Makers of Mars Observer, by contrast,faced no such constraint. Scheduled to launch ona much larger Titan III, they allowed the size oftheir spacecraft to growand the cost of themission increased commensurately. Confining

missions to small launchers created a physicalenforcement mechanism over cost and mass thatteams developing faster, better, cheaper space-craft could not avoid.

Controlling RiskIn promoting the faster, better, cheaper ini-

tiative, Dan Goldin told his spacecraft managers totake more risks. Be bold, he said, push the lim-its of technology.15 By that, he meant that man-agers should utilize new technologies as a means ofimproving the performance of spacecraft. He didnot mean that managers should trade away per-formance in their effort to reduce the cost and timerequired to prepare spacecraft for flight.

Essentially, all of APLs space missions areone-of-a-kind, observed two senior managersfrom the APL Space Department. So the challengeis to get it right the first time. In each case, space-craft managers utilize technology to improve per-

The NEAR spacecraft team tooused solar panels to power the spacecthe fact that no spacecraft using soleven flown that far from the Sun. Th

install a highly sophisticated mechanimovement in the spacecraft, called resonator gyroscopes, which utilize a pattern on a piece of quartz rather ttional spinning wheel. No spacecrflown with such a gyroscope.17 In risks, the NEAR team was much less wthe reliability of their spacecraft than group, which took extensive risks a

minds of NEAR team members, ththrough this before. Nonetheless, tmany precautions as a means of coupotential for difficulties arising from gies they used.

Team members reworked the spacWhen early design studies revealed thsilicon solar cells would not produce

tric power, the NEAR team switcharsenide. Team members built in redspacecraft carried three separate attmechanismsa star camera, a fullinertial guidance unit, and redundasensors.18 The inertial guidance utwo power supplies, two processorand four accelerometers.19 The spacthree methods of communicating with

gain antenna, two low-gain antennas aft), and a fanbeam antenna with twpatch arrays especially important in tthe spacecraft during emergency situat

Where the design team participaimpart redundancy, they built in mrisk with margin is a key strategy foreliability into complex systems. Ma

the tolerance for miscalculations, prcraft for new environments, and prunanticipated errors.

The NEAR team built margins incomponents. Doing so was especially


34/86

that a spacecraft on which he had previouslyworked (ISEE-3) carry extra fuel allowed it to makean unplanned rendezvous with comet Giacobini-Zinner.21 The team put extra fuel on NEAR.

Redundancy and larger margins worked toincrease the cost of the spacecraft. They also madethe spacecraft more complex. The people designingthe spacecraft balanced increased margins andredundancy with simplification in other areas.Complexity and simplicity coexisted within thesame machine. The NEAR team simplified thedesign of the overall spacecraftespecially as they

attached components like the antennas and solarpanels to the body of the machineand theysimplified the interfaces between spacecraft compo-nents. Interfaces are the points at which compo-nents join and are the most common source ofcomplex system failure. The propulsion system wasdesigned as a separate module to simplify thepropulsion system-to-spacecraft interfaces andgreatly reduce schedule risk, said propulsion engi-

neers.22 Simplification of many components coun-terbalanced the risks taken with others.

The short and unyielding work schedule alsopromoted spacecraft simplicity because it acted as abarrier against mission creep. This, too, helpedto reduce cost and risk. Excessively long develop-ment schedules create windows of opportunity thatengineers may fill by increasing the technical com-

plexity of a spacecraft. Although excessive tinkeringmay be done with the intent of reducing risk, itoften has the opposite effect, as the history of theMars Observer mission demonstrated. The NEARteam, faced with an unusually tight developmentschedule, had to keep the spacecraft simple. Theteam was obliged to use engineering methods thatwere more linear than interactive in nature. Linearmethods, or what engineers call clean interfaces,

enhance reliability. With their February 1996launch window approaching, the NEAR team couldnot engage in actions that made the spacecraft toocomplex.

Redundancy, margins, simplicity, and limited

anced, as was the case with the NEAReffort, the contest between reliabilitoften decided by the quality of the team. Here, as well, great cost saving

Managing for None of the money that spacecr

consume is spent in space. All of it puand services produced on Earth, andportion of it pays the salaries of peop

the project. Spacecraft projects are lawith far more funds devoted to salhardware. During the 1990s, the expense of a single employee workinproject averaged $95,000, including fits, and administrative overhead.23 Anlasts twice as long as usual and emploas many people will, of necessity, cosmore than one that employs a smaller

a shorter development period.

The practice of using small, rapidefforts was pioneered by Clarence Kea Lockheed Corporation engineer produce Americas first production-liduring World War II. Johnson set works development team that wasmall for its task and assigned a

responsibility to individual team memembers were isolated from ongoing that they could concentrate on theirAccording to one version, the skunkarose from the relocation of the teamremoved from the main corporate uncomfortably close to an adjacent pthat emitted a disagreeable smell.

Aerospace managers applied thphilosophy embodied in the skapproach to relatively small, focuScientists at APL utilized small teambegan conducting experiments with soets after World War II; they continue


35/86

Shortly after the age of modern rocketry began,managers overseeing the largest undertakingsadopted a different approach. Confronted with the

used to develop the first fleet of U.S. tal ballistic missiles and to completMoon project. Given the complexity o

Kelly Johnson Skunk Works Rules

The skunk works manager must be delegated practically complete control of h

program. The manager should report to a division president or higher.

Strong but small project offices must be used by everyone involved.

The number of people having any connection with the project must be restrictalmost vicious manner. Use a small number of good people.

A very simple drawing and drawing release system with great flexibility for machanges must be provided in order to recover from failures.

There must be a minimum number of reports required, but important work murecorded thoroughly.

There must be a monthly cost review covering not only what has been spent ancommitted but also projected costs to the conclusion of the program. Dont suthe customer with sudden overruns.

Team members must be delegated and must assume more than normal responto get good vendor bids for subcontracts on the project.

Push basic inspection responsibility back to subcontractors and vendors.

Team members must be delegated the authority to test their final product in flThey can and must test it in the initial stages.

The specifications applying to hardware must be agreed to in advance of contr

Project funding must be timely so that the skunk works manager doesnt haverunning to the bank to manage cash flow.

There must be mutual trust between the customer and the skunk works team wclose cooperation and liaison on a day-to-day basis. This cuts down misunderstand correspondence to an absolute minimum.

Access by outsiders to the project and its personnel must be strictly controlled

Because only a few good people will be used, ways must be provided to pay pebased on performance and not on the number of people supervised.

Source: Ben R. Rich and Leo Janos, Skunk Works (Boston: Little, Brown, and Company, 1994), pp. 5153.


36/86

Large-scale systems management consists of aseries of formal techniques designed to give projectleaders ever-increasing degrees of control over thedesign, financing, and reliability of machinery

under development. Progressive design freezes fixthe configuration of spacecraft and related subsys-tems at increasingly specific levels of detail. Formalscheduling techniques (such as the critical pathmethod, or CPM) track the completion of individ-ual tasks. Total program budgets and multiyearfinancial plans track spending. Formal programreviews provide management oversight and identifypressing difficulties. Quality and reliability units,

located outside the project hierarchy, police thework of project managers while project managerspolice the work of contractors. At the top, formallyorganized systems engineering and integration unitsconduct studies to determine the potential for inter-active failures arising from the simultaneous opera-tion of two or more subsystems. The integrationunits also check the effect of small design changeson adjacent machinery.

Formal, full-scale systems management is time-consuming, labor-intensive, and expensive. Itresults in an elaborate network of checks and bal-ances wherein the result of any one action isassessed for its effect on the spacecraft, the projectbudget, and the development schedule. Leaders ofinstitutions like NASA utilize formal systems man-agement to reduce the probability of unanticipated,

interactive failures once the mission is under way.Designing and building large spacecraft without theadvantages provided by formal systems manage-ment is very riskythe equivalent of conducting acircus trapeze act that begins its performance with-out a net. The smallest undetectable error can doomthe final result.

Faster, better, cheaper projects typically

forgo the more excessive demands of formal sys-tems management, substituting instead the advan-tages that arise when members of a small teamintensively focus on a specific objective and feel per-sonally responsible for the collective product oftheir individual contributions. Even before people

with each other and coordinate thcontributions in an informal sort of w

Leaders of both the NEAR and t

projects relied much more on teamwformal systems management to carwork. In essence, they substituted tformal systems management. Howeveabandon formal systems managementNEAR team, for example, engaged inof technical, cost, design, readiness, qsubsystem reviews. Like the Pathfindestablished peer-review groups befor

members had to defend their work. Thgroups drew representation from insDepartment, from other units withfrom people outside APL. Thomas NEAR project manager, established thbility and quality-assurance componenpanies formal systems management the APL practice of vesting this respoa product-assurance engineer who sat

of the core team.

Coughlin departed from formal agement by avoiding the excessive paporate integration studies, and mreviews that characterize that approacthat NASA executives, the people pmission, establish a single point of cosee APL. When project teams are force

numerous supervisors and multiple Fthe small alterations in mission requtend to arise create huge cost and scalies.25 Years earlier, Richard B. Kfounder of APLs Space Department, htake any further NASA contracts befrustrated over the requirement thaengineers squander every Friday NASA oversight counterparts.26 To

single NASA Headquarters represewhom he worked seemed worried in tShe did not know whether the APL tthe work. Still, she protected the teamNASA officials who wanted to arequirements and add more instrumen


37/86

the technical people and lead engineers to attend.The meetings lasted 1 hour. No coffee was served.Each person had 3 minutes to report.

By eliminating the formal elements of large-scale systems management, Coughlin kept theNEAR project team small. It had to be kept small.A large team, with heavy paperwork requirements,never could have met the impending spacecraftdevelopment deadlines. The February 1996 launchwindow had the same effect on Coughlins projectmanagement methods that the Delta 2 rocket hadon spacecraft size. Both kept their respective prod-

ucts small. The impending launch date preventedother people from adding new requirements thatwould make the project slow and expensive. Youhave to shut everything out, Coughlin remarked.That [launch window] was the biggest friend I had.28

The core team at its height consisted of perhaps40 people. This was quite small by spacecraft stan-dards. Managers of large planetary missions com-

monly employ thousands of people to develop theirspacecraft. Even for small projects, the core devel-opment team can exceed 100 employees. Coughlinestimated that the whole number of people workingto develop the project, including people in the APLfabrication shops, totaled no more than 130 APLemployees, with perhaps another 170 doing workthat had been contracted out. We tried to mini-mize the [number of] people who touched the hard-

ware,29 Coughlin explained. The decision chainwas so short that people on the team could makemajor decisions in a day or so.

Managers of low-cost projects at the AppliedPhysics Lab, like those working at the JetPropulsion Laboratory, use an organizationalapproach called matrix management. Workers aredrawn into the project organization from other

units without formally removing them from theirhome units and the supervisors to whom theyreport. The matrix refers to the way in which theproject team is superimposed over the formalorganization chart. APL consists of departmentswhere project workers, technical experts, and man-

fabrication shops, where hardware is bThe overall organization is further cothe involvement of scientists from uother research laboratories who ov

ation of instruments and the involvetractors who produce spacecraft partmanagement system consists of a smproject leaders, reporting directly toment heads, drawing help from a muwork of sections, groups, branches, a

Effective teamwork is essenapproach. In the absence of the form

vided by full-scale systems managemlines of authority created by a strict mand, project leaders must rely uponess of participants to develop a perresponsibility for mission success andconcerns openly with one another. leaders must rely upon the capacity oto function as a team. Like the Pathand other successful faster, better, c

ects, NEAR team leaders utilized a nuniques that enhanced this capacity fo

NEAR leaders colocated team mAPL campus in central Maryland. a primary technique for promoting of interpersonal communication nteamwork to occur. It is especially impcritical systems integration function. D

systems integration function among sites, especially on a low-cost project, failure. Colocation of team memberssafeguard against the communication darise on complex undertakings.

Most of the work on the actual sdone in-house at APL. Hands-on exthe spacecraft gives team members

with hardware that cannot be crepaperwork reviews. APL had a tradpleting 70 percent of any project in-hof its short development schedule, thecraft team retained 50 percent of its wing out components such as the


38/86

who could understand how their contribution tothe project fit into the system as a whole. Managersconducting low-cost projects often use a techniquecalled multitasking, giving workers more than onejob to perform so that they view the project fromdifferent perspectives, thus expanding individualunderstanding of the overall machine. Project lead-

ers also encourage seamless management, wherebythe same people move from design to fabricationto operational tasks. Coughlin appointed leadengineers for each subsystem, working out theirassignments with the supervisors of the units fromwhich these persons were drawn. Having one

entific and technical objectives of the single highest risk to managing a cosgram is requirements creep, CouThis is it, he said, fighting designpointing to the 1996 launch windowextra time to absorb design changestime in the later stages of developmen

a standing army is being extended, expensive and also tends to defocus th

The exceptionally short developmhelped Coughlin keep people on the teaprojects, as design and fabrication eff

The organization established to design and build the low-cost spacecraft consisted of a small APL team assisted by sciefor the instruments the spacecraft would carry.



39/86

Overall, Coughlin used a relatively simple proj-ect organization. He oversaw a small number ofpeople responsible for the spacecraft, its scientificinstruments, mission design, and in-flight opera-

tions. He organized six small science teams to assistwith the fabrication and operation of instruments,with each team headed by an outside expert whosework was counterbalanced by instrument special-ists selected from APL. Most of the fabricationwork was done in-house at the Maryland campusor by the Aerojet Propulsion Company, whichproduced the propulsion subsystem.

Writing about the experience, Coughlin sum-marized the management principles that guided thework of the project team. In his typically modestway, Coughlin credited the principles to Richard B.Kershner, who founded the APL Space Departmentin 1959 and served as its head for nearly 20 years.

John Dassoulas, an APL program manager andproject engineer, codified Kershners ideas in severalpublications.32

Kershner was a brilliant mathematician whoplaced more trust in the problem-solving capabilityof individuals than in the coordinating powers offormal management systems. Emphasizing individ-ual skill, Kershner observed:

In carrying out a successful developmentprogram, the importance of recognizing

individual differences in staff cannot beoveremphasized. It is precisely this recogni-tion of differences that is least stressed inmany books and articles on the theory oforganization and management.33

In modern terms, Kershner was more interestedin leadership than in management. A colleaguenoted that he always made it fun, and could get a

bunch of screw-balls to work together.34

Three of Kershners seven management guide-lines for low-cost missions were designed to build atight-knit team that would stay focused on the proj-ect. Establish a small team of experienced person-

required to inform new personnel abalready done. Two of the guidelines

challenge of restraining project spendto use cost as a design parameter anment that reliability and redundancinto the mission at the start of the pto build in reliability once a project wCoughlin observed, invariably caused One guideline dealt with the position assigned to check on reliability and the practice in large programs where

sit outside the project organization, preferred to appoint a product-assuras part of the development team. The addressed the method of projeemployed by the customer or the insmissioning the project. APL officials in

Table 3-5

APL Management Guideli

for Low-Cost Missions1. Limit the schedule from start to l

months or less.

2. Establish a small, experienced tec

3. Design the spacecraft and instrum

4. Use the lead-engineer method for ea

5. Design in reliability and redundanc

6. Integrate the product-assurance ethe program.

7. Assign a single Agency manager with the development team.

Source: Thomas B. Coughlin, Mary C. CDassoulas, Forty Years of Space Mission ManHopkins APL Technical Digest 20 (OctoberD507509.


40/86

ment that major subsystems be able to stand alone.He encouraged spacecraft engineers to attend themeetings of teams designing instruments. He keptthe project team small, maintained in-house techni-cal capability, and balanced the need for outsourc-ing with the necessity of completing at least half ofthe fabrication work at APL. He insisted that the

lead engineers and the team as a whole feel person-ally responsible for the success of the mission anddefined mission success to include cost and scheduleas well as scientific goals.

Repeatedly, Coughlin and others on his team

Development of the NEAR spacectively uneventful by comparison to othtype missions like Viking and Mars Oproject team held its preliminary desApril 1994, four months after fundiflow. The team completed its critical the following November. Testing of in

ponents and assembly of the spacecJune (1995), following the arrival of tsubsystem from the Aerojet Companyscale testing began in the last week of the Maryland lab, technicians subjeccraft, in its nearly complete flight con

Pursuing a fast development schedule, members of the NEAR project team began full-scale testing of the assembled 1995, less than 24 months after mission funding began.



41/86

low cost innovation in spaceflight the history of the near earth asteroid rendezvous (near) mission

Documents