Affordable Spacecraft: Design and Launch Alternatives · IBM Federal Systems Division George B....

Affordable Spacecraft: Design and LaunchAlternatives

January 1990

OTA-BP-ISC-60NTIS order #PB90-203225

Recommended Citation:

U.S. Congress, Office of Technology Assessment, Affordable Spacecraft: Design and LaunchAlternative --Background Paper, OTA-BP-ISC-60 (Washington, DC: U.S. GovernmentPrinting Office, January 1990).

Library of Congress Catalog Card Number 89-600776

For sale by the Superintendent of DocumentsU.S. Government Printing Office, Washington, DC 20402-9325

(order form can be found in the back of this report)

Foreword

Several major DOD and NASA programs are seeking ways to reduce the costs oflaunching spacecraft. However, it typically costs much, much more to build a spacecraft thanto launch it into a low orbit. Therefore, unless the costs of building spacecraft are reduced, evendramatic reductions in costs of launching to low orbit would reduce total spacecraft programcosts by only a few percent.

This background paper examines several proposals for reducing the costs of spacecraftand other payloads and describes launch systems for implementing them. It is one of a seriesof products of a broad assessment of space transportation technologies undertaken by OTAat the request of the Senate Committee on Commerce, Science, and Transportation, and theHouse Committee on Science, Space, and Technology. In 1988, OTA published the specialreport Launch Options for the Future: A Buyer’s Guide and the technical memorandumReducing Launch Operations Costs: New Technologies and Practices. In 1989, OTApublished the background paper Big, Dumb Boosters: A Low-Cost Space TransportationOption? and the special report Round Trip to Orbit: Human Spaceflight Alternatives. Asummary report on space transportation, entitled Access to Space to be published in the springof 1990, will be the final report in the space transportation series.

In undertaking this effort, OTA sought the contributions of a wide spectrum ofknowledgeable individuals and organizations. Some provided information; others revieweddrafts. OTA gratefully acknowledges their contributions of time and intellectual effort. OTAalso appreciates the cooperation and assistance of the Air Force and NASA. However, OTAis solely responsible for the content of this background paper and the other OTA publications.

u JOHN H.-GIBBONSDirector

Advisory Panel on Affordable Spacecraft:Design and Launch Alternatives

M. Granger Morgan, ChairHead, Department of Engineering and Public Policy

Carnegie-Mellon University

Michael A. BertaAssistant Division ManagerSpace and Communications GroupHughes Aircraft Company

Richard E. BrackeenPresidentMartin Marietta Commercial Titan, Inc.

Edward T GerryPresidentW. J. Schafer Associates, Inc.

Jerry GreyDirector, Science and Technology PolicyAmerican Institute of Aeronautics and Astronautics

William H. HeiserConsultant

Otto W. Hoernig, Jr.Vice PresidentContel/American Satellite Corporation

Donald B. JacobsVice President, Space Systems DivisionBoeing Aerospace Company

John LogsdonDirector, Space Policy InstituteGeorge Washington University

Hugh F. LowethConsultant

Anthony J. MacinaProgram ManagerIBM Federal Systems Division

George B. MerrickVice PresidentNorth American Space OperationsRockwell International Corporation

Alan ParkerSenior Vice PresidentTechnical Applications, Inc.

Gerard PielChairman EmeritusScientific American

Bryce Poe, IIGeneral, USAF (retired)Consultant

Ben R. RichVice President and General ManagerLockheed Corporation

Sally K. RideCalifornia Space Institute

Tom RogersPresidentThe Sophron Foundation

Richard G. SmithSenior Vice PresidentJLC Aerospace Corporation

William ZersenConsultant

NOTE: OTA appreciates the valuable assistance and thoughtful critiques provided by the advisory panel members. The viewsexpressed in this OTA report, however, are the sole responsibility of the Office of Technology Assessment. Participation onthe advisory panel does not imply endorsement of the report.

iv

OTA Project Staff onAFFORDABLE SPACECRAFT—DESIGN AND LAUNCH ALTERNATIVES

Lionel S. Johns, Assistant Director, OTAEnergy, Materials, and International Security Division

Alan Shaw, International Security and Commerce Program Manager (from March 1989)

Peter Sharfman, International Security and Commerce Program Manager (through February 1989)

Ray A. Williamson, Project Director (from February 1989)

Richard DalBello, Project Director (through January 1989)

Michael B. Callaham, Principal Analyst

Eric O. Basques (through August 1989)

Gordon Law

Administrative Staff

Jackie Robinson Louise Staley

Acknowledgments

The following organizations generously provided OTA with information and suggestions:

American Rocket Corporation

Boeing Aerospace Company

California Institute of Technology,Jet Propulsion Laboratory

Defense Advanced Research Projects Agency

Defense Systems, Inc.

Electromagnetic Launch Research, Inc.

NASA Headquarters

Naval Research Laboratory

Orbital Sciences Corporation

Radio Amateur Satellite Corp.

Strategic Defense Initiative Organization

TRW, Inc.

U.S. Air Force Systems Command, Space SystemsDivision

U.S. Space Command

This paper has also benefited from the advice of many experts from the Government and the privatesector. OTA especially would like to thank the following indivuals for their assistance and support. The viewsexpressed in this paper, however, are the sole responsibility of the Office of Technology Assessment.

Steve AltesOrbital Sciences Corporation

Ivan BekeyNASA Headquarters

James BennettAmerican Rocket Corporation

Carl BuilderThe Rand Corp.

Paul Dergarabedian, Sr.Aerospace Corp.

George DonohueThe Rand Corp.

Dani EderBoeing Aerospace Company

John GainesGeneral Dynamics

Daniel GregoryBoeing Aerospace Company

Col. Charles Heimach, USAFUSAF Systems Command, Space

Systems Division

Abraham HertzbergUniversity of Washington

Ross M. JonesCalifornia Institute of TechnologyJet Propulsion Laboratory

Jordin KareStrategic Defense Initiative

Organization

Clark KirbyTRW, Inc.

Henry KolmElectromagnetic Launch Research,

Inc.

George KoopmanAmerican Rocket Corporation

Robert LindbergOrbital Sciences Corporation

Andrew C. McAllisterRadio Amateur Satellite Corp.

Lt. Col. Tom O’Brien, USAFU.S. Space Command

Scott PaceU.S. Department of Commerce

Miles PalmerScience Applications International

Corp.

David RossiOrbital Sciences Corporation

Arthur Schnitt

George SebestyenDefense Systems, Inc.

Larry SternGeorge Mason University

Peter WilhelmNaval Research Laboratory

Col. John Wormington, USAFUSAF Systems Command, Space

Systems Division

vi

ContentsPage

Chapter 1. Summary ..o. ..o. ... o.. ... ... ... .*. ... *.. ... ... ... ... Q.*...**""""""*""""""""""""" 1THE HIGH COST OF SPACECRAFT’ . . . . . . . . . . . . . . ........................................s 1APPROACHES TO REDUCING PAYLOAD COSTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2OPTIONS FOR CONGRESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Options for Influencing Spacecraft Design . . . . . . . . . . . . . . . . . . . . . . . . . .. .. .. .. .. .. ... .... ....O. 4Options for Promoting the Development of Launch Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Chapter 2. Weight and Volume Margins ● ....0.0 . . . . . . . . . . . . . . . . . . . . . . . . . . . ● .00..00. ● ...0...0 7

Chapter 3. Fatsats . . . . . 0 0 0 . . . . . 0 0 0 . . . . . . . . . . . . . . . . . . . . . 0 . 0 . 0 . 0 . 0 . . . . . 0 0 . 0 . . 0 . 0 . . . 0 0 . 0 . . . . . 0 0 9STANDARDIZING SUBSYSTEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9UPPER STAGEHAND ORBITAL-TRANSFER VEHICLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10ESTIMATING POTENTIAL COST REDUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

A Parametric Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12Bottom-Up Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14A Comparison of Parametric and Bottom-Up Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

ACHIEVING POTENTIAL COST REDUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Chapter 4. Lightsats ......00 . . . . . . . . . . . . . . . . . . .... . . . . . . . ....000.0 ●19

SPACECRAFT CONCEPTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19CONCEPTS FOR PHASED ARRAYS OF SPACECRAFT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20LAUNCH REQUIREMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22LAUNCH SYSTEM OPTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23

Existing Launch Vehicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Air-Launched Vehicles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23Standard Small Launch Vehicle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25Other Options.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

ISSUES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

Chapter 5. Microspacecraft . 0 . 0 0 . . 0 ● . 0 0 . . . . . . . . . . . . . . . . . 0 0 . 0 . . ● . . . . . 0 0 0 ● . . . . . . . 0 ● 0 . . . 0 0 0 . . . . . 29SPACECRAFT CONCEPTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29LAUNCH SYSTEM OPTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29ISSUES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

BoxesBox Page3-A. Fatsats: The Good and Ugly, and the Bad . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174-A. Brilliant Eyes? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 285-A. Lasers for Rocket Propulsion: The State of the Art . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

FiguresFigure Page2-1. Value of Weight Margin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73-1. Average Launch Cost v. Payload Capability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133-2. Cost for Hypothetical Mission With Current Launch Cost Trend . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133-3. Cost for Hypothetical Mission If Launch Costs Are Reduced 67 Percent . . . . . . . . . . . . . . . . . . . . . . . 133-4. Payload Cost v. Weight Trade-offs and Economic Launch Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143-5. If a Satellite Were Allowed To Be Heavier It Could Cost LESS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153-6. The Effect of Reducing Launch Cost on the Optimal Weight and Cost of a Payload. . . . . . . . . . . . . 154-1. The Global Low-Orbit Message Relay (GLOMR) Satellite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204-2. Microsat: A 22-lb Communications Satellite for Amateur (“Ham’’)Radio Operators . . . . . . . . . . . . 214-3. Sparse Array of Satellites Could Provide Detailed Radar Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 224-4. Orbiting Low-Frequency Array of Radioastronomy Satellites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234-5. Pegasus Launch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244-6. Projected Performance of Pegasus: Payload v. Orbital Altitude and Inclination . . . . . . . . . . . . . . . . . . 255-1. Laser-Powered Rockets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305-2. Mars Observer Camera Microspacecraft Design Cutaway View.... . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

TablesTable Page3-1. Ratios of Nonrecurring to Recurring Cost of Spacecraft Subsystems . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 05-1. Estimated Cost of a 20-Megawatt Laser for Powering Rockets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

vii

Chapter 1

Summary

Several major efforts are aimed at findingways to reduce the cost of launching spacecraft.However, it typically costs much, much more tobuild a spacecraft than to launch it to low Earthorbit (LEO). Unless spacecraft costs are re-duced, even dramatic reductions in launch costswill have only a small effect on total spacecraftprogram costs.

This Background Paper reviews four possibleapproaches to spacecraft design that have beenproposed to reduce total spacecraft programcosts. Adopting them could change the launchrates, payload capacity, and reliability demandedof conventional launch vehicles, or create ademand for exotic launch systems to launch verysmall spacecraft cheaply. Conversely, develop-ing new, economical launch systems wouldstrengthen incentives to adopt these new ap-proaches to spacecraft design.

This is one of several publications document-ing OTA’s broad assessment of space transporta-tion technologies requested by the House Com-mittee on Science, Space, and Technology, andthe Senate Committee on Commerce, Science,and Transportation. Previous publications inthis assessment examined a variety of futurelaunch options,l ways to reduce the costs oflaunch operations,2 low-cost, low-technology(“big, dumb") boosters,3 and options for trans-porting humans to and from orbit.4 A final reportwill be published in 1990.

THE HIGH COST OFSPACECRAFT

Because of the high cost of spacecraft, adramatic reduction in launch cost alone willnot substantially lower spacecraft programcosts. Although launching a pound of payload toLEO currently costs about $3,000, procuringthat pound of payload typically costs muchmore. For example, representative U.S. space-craft bussess of types first launched between1963 and 1978 cost between $130,000 and$520,000 per pound dry,6including amortizedprogram overhead costs. Procurement of themission payloads carried on those busses costabout 50 percent more—about $200,000 to$800,000 per pound.7 Reducing launch costsfrom $3,000 to $300 per pound of payload, agoal of the Advanced Launch System program,8

would reduce the total cost of procuring andlaunching a dry spacecraft (half bus, halfmission payload) by less than 2 percent.

A spacecraft bound for a high orbit or anotherplanet requires an upper stage, which whenfueled is typically more than twice as heavy asthe spacecraft but costs less. Even so, a payloadconsisting of a Centaur upper stage (about$2,250 per pound) and a spacecraft weighing athird as much (half bus, half mission payload)might cost from $40,000 to $160,000 per pound.Reducing launch costs to $300 per pound would

Iu.S. Conm55, office of l’whnolo~ Assessment, Launch Options for the Future: A Buyer’s Guide, OTA-ISC-383 (Washington, DC: U.S.Government Printing Office, July 1988).

2u.s. Conge55, office of T~hnoloW Assessment, lled~ing hunch Operations Costs: New Technologies Und practices, OTA-TM-ISC-28(Washington, DC: U.S. Government Printing Office, September 1988).

3u.s. conw~5, office of TWhnology As~ssment, Big, D@ Bo osters : A ~w-cost space Tra~pO~~ion option? (WfaSMn~On, DC: U.S.Congress, Office of Technology Assessment, International Security and Commerce program, February 1989).

4u.s. con=e55, office of T’ec~oloW A~ssment, Rou~ Trip ~ orbit: H~an spuceflightAfter~tives, OTA-ISC-419 (Washington, DC: U.S.Government Printing Office, August 1989).

5~e c(bw~~ of a ~acar~t consists of the s~ctme, ~wer ~Wce5, and o~er subsys~ms r~uired to support the rnisslon payloads it carries.

%.Jnfueled.7Space System ad Operatlom Cost Reduction and Cost Credibility Workshop, Executive S ummary (Washington, DC: National Security

Industrial Association, 1987), p. 3-10, fig. 3.7.3. These estimates were derived by amortizing nonrecurring costs (e.g., for development) over foursatellites; some programs procure more than four while others procure only one. OTA inflated the estimated costs, which were in 1982 dollars, to1987 dollars using the GNP inflator from table B-3 of the ECOrWndC Report ofthe President, January 1989, and then to 1989 dollars by assuming 4.2percent annual inflation in 1987 and 1988.

8101 s~t. 1067.

–l–

2 ● Affordable Spacecraft: Design and Launch Alternatives

reduce the total cost of procuring and launchingsuch a payload by only 2 to 6 percent.

APPROACHES TO REDUCINGPAYLOAD COSTS

To reduce payload costs, and for otherreasons, novel approaches to payload design andfabrication have been proposed:

● Design payloads to fit launch vehiclesleaving size and weight margins of about15 percent

. Allow payloads to be larger and heavier:Fatsats

● Allow satellites to be simpler, and makethem lighter: Lightsats

● Design Microspacecraft to be launched likeartillery shells

Each type of spacecraft—fatsat, lightsat, ormicrospacecraft-would impose unique launchdemands. New, large launch vehicles would beneeded to launch the heaviest satellites.Lightsats could be launched on existing launchvehicles, but new, smaller launch vehicles mightlaunch them more economically. In wartime,small launch vehicles could be transported orlaunched by trucks or aircraft to provide asurvivable means of space launch. Microspace-craft could be launched on existing launchvehicles, but they might instead be launched bymore exotic means such as a ram cannon,railgun, coilgun, or laser-powered rocket. Someof these might be proven feasible in the nextdecade.

Weight margin: Designing payloads to fitlaunch vehicles while reserving ample sizeand weight margins can reduce the risk ofincurring delay and expense after assemblyhas begun.

It is often the case that satellites growsubstantially heavier than expected as theyproceed from design to construction. For ex-

ample, dry weights of military spacecraft havebeen about 25 percent greater, on the average,than initially predicted. Growth in estimatedweight may be caused by underestimating theweight of a spacecraft (especially one designedto use the most advanced technology) or bychanging mission requirements during develop-ment (requiring hardware to be added). If apayload grows so heavy during assembly that itthreatens to “gross out” its assigned launchvehicle (i.e., cause its weight to equal or exceedthe maximum allowable gross lift-off weight),the payload must be redesigned to cut its weight.This causes delay and increases cost.

To reduce the risk of exceeding vehiclepayload capacity, program managers could re-quire designers to design each payload to fit itsassigned launch vehicle with room to spare andto weigh substantially less than the maximumweight the vehicle can launch to orbit. However,this design philosophy would lead to morestringent size and weight constraints than wouldotherwise be imposed. If a mission simply couldnot be performed by payloads predicted to besmall enough to fit the largest launch vehicleswith adequate size and weight margins, new,larger launch vehicles would have to be devel-oped to provide the desired margin. In manycases, however, sufficient margin could beprovided by clever design, e.g., by designingseveral smaller single-mission payloads insteadof a single multimission payload, or developingand using an electric-powered or space-basedorbital transfer vehicle (OTV) instead of aconventional OTV.9

Fatsats: If payloads were allowed to beheavier for the same capability, some couldcost substantially less. For example, OTAestimates that Titan-class payloads that costseveral hundred million dollars might costabout $130 million less if allowed to be fivetimes as heavy.

9To place a satellite in a high orbit, a launch vehicle must carry both the satellite and either an upper stage to take the satellite directly to the highorbit, or an OTV to take the satellite from a low-altitude parking orbit to the high orbit. A conventional OTV weighs two or three times as much as thesatellite it carries. An electric-powered OTV could weigh less than a conventional OTV of comparable capability, creating more weight margin. Aspace-based OTV could be launched separately from its payload, allowing the payload to weigh as much as its launch vehicle could carry while reservingthe desired weight margin. However, operation of space-based OTVs would be complex and require costly infrastructure.

Chapter I-Summary ● 3

If payloads were allowed to be much heavier,a manufacturer could forego expensive proc-esses for removing nonessential structural mate-rial, as well as expensive analyses and tests forassuring the adequacy of the remaining struc-ture. Standardized subsystems, which could beproduced economically in quantity, could beused instead of customized subsystems de-signed to weigh less.

The savings that might be realized are uncer-tain. In principle, they could be estimated bycomparing the costs of a heavy payload and alight payload that perform the same functionswith the same capability. However, the UnitedStates has never designed and built two pay-loads, one heavy and the other light, thatperform the same functions equally well, inorder to compare actual costs. A few estimateshave been derived by comparing the cost andweight of an actual spacecraft with the estimatedcost and weight of hypothetical heavier space-craft of comparable capability. Designers havealso compared the estimated costs and weightsof hypothetical spacecraft of comparable capa-bility and different weights. All such studiespredict payloads could cost less if allowed toweigh more, but the estimates of savings differ.

An accurate estimate of potential savingsrequires a detailed trade-off analysis for eachpayload. Achieving these savings will probablyrequire giving spacecraft program managers,and those who establish mission and spacecraftrequirements, incentives crafted specifically forthe purpose, and may require developing newlaunch or orbital-transfer vehicles to carry thespacecraft.

Lightsats: If allowed to be less capable,reliable, or long-lived, payloads could be bothlighter and less expensive. Useful functionssuch as communications and weather surveil-lance could be performed by payloads smallenough to be launched on small rockets fromairborne or transportable launchers.

Small, simple, and relatively inexpensivecivil and military satellites have been, and stillare, launched at relatively low cost on smalllaunch vehicles or at even lower cost, sometimesfor free, as “piggyback” payloads on largerlaunch vehicles.

The Department of Defense is consideringwhether the increased survivability and respon-siveness such spacecraft could provide wouldcompensate for possibly decreased capability.Some missions might be accomplished as wellby a swarm of several small satellites as by asingle large one. If so, a swarm would be lessexpensive in many cases, because smaller satel-lites typically cost much less per pound than dolarge ones. Even if the satellites were launchedindividually, which would increase total launchcost, total mission cost might be lower.

Microspacecraft: Spacecraft weighing onlya few pounds could perform useful spacescience missions and might be uniquely eco-nomical for experiments requiring simulta-neous measurements (e.g., of solar wind) atmany widely separated points about theEarth, another planet, or the Sun.

These could be launched on existing launchvehicles. Eventually, it may be possible tolaunch them on laser-powered rockets or, if theyare as rugged as a cannon-launched guidedprojectile, 10 with a ram cannon or an electro-

magnetic launcher. Within the next decade,experiments now being planned may establishthe feasibility of some of these launch systems.Their costs cannot be estimated confidentlyuntil feasibility is proven, but at high launchrates they might be more economical thanconventional rockets. An electromagnetic launchercould also be constructed in orbit to launchmicrospaceprobes to outer planets; they wouldarrive years earlier than if they were propelledby conventional rockets.

IOA camon-lauc~ guided projectile is an artillery shell equipped with a system (e.g. TV camera and computer) for recomztig a target admovable fm or other means for steering the projectile toward a target. The Army’s M712 Copperhead is an example.


OPTIONS FOR CONGRESSWhat can Congress do to promote spacecraft

cost reduction and, thereby, reduce the cost ofspace programs? Some options deal directlywith spacecraft design; others would promotethe development of launch systems that couldlaunch small, inexpensive spacecraft at low costor heavy spacecraft with generous weight mar-gins.

Options for Influencing Spacecraft Design

Option 1:

Congress could order a comprehensive studyof how much the Nation could save on spaceprograms by:

●

●

●

●

●

designing payloads to reserve more weightand volume margin on a launch vehicle;allowing payloads to be heavier, lesscapable, shorter-lived, or less reliable;designing standard subsystems and busesfor use in a variety of spacecraft;designing spacecraft to perform singlerather than multiple missions; andusing several inexpensive satellites insteadof a single expensive one.

Lockheed completed such a study in 1972;11

anew one should consider current mission needsand technology. It would complement the SpaceTransportation Architecture Study (STAS) andmore recent and ongoing studies12 that comparespace transportation options but not payloaddesign options.

As noted above, to estimate potential savingsaccurately, a detailed trade-off analysis must bedone for each payload, or more generally, foreach mission. So, for greater credibility:

Option 2:

Congress could require selected spacecraftprograms--for example, those that might re-quire a new launch vehicle to be developed--toaward two design contracts, one to a contractorwho would consider the unconventional ap-proaches mentioned above.

Option 3:

Congress could require both the Departmentof Defense and NASA to refrain from developinga spacecraft if the expected weight or size of thespacecraft, together with its propellants, upperstage, and support equipment, would exceedsome fraction of the maximum weight or sizethat its intended launch vehicle can accommo-date. Public Law 100-456 required the De-partment of Defense to require at least 15percent weight margin in fiscal year 1989.13

New legislation could extend this restriction toNASA and could require size margins in futureyears.

Options for Promoting the Development ofLaunch Systems

Option 4:

Congress could fund the development of theShuttle-C cargo launch vehicle, the AdvancedLaunch System, liquid-fueled rocket boostersfor the Shuttle, or a larger Titan launchvehicle. 14 Any of these vehicles could launchpayloads larger and thus less expensive (forcomparable performance) than payloads de-signed to fly on Shuttles or Titan IVs. Alterna-tively, if payload size and weight are notincreased, these proposed launch vehicles couldprovide greater size and weight margins, therebyreducing the risk of needing costly weight-reduction efforts. However, their greater pay-load capacity would also enable payload pro-—

ll~ckh~ ~ssiles and Space CO., ]Wact of Low-Cost Rejivbishable and Standard Spacecraji Upon Future NASA Space Progmms, NTISN72-27913, Apr. 30, 1972.

%g., tie Air Force’s Air Force-Foc~ed SEAS, NASA’s Next Manned Space Transportation System study, the Defense Science Board’s NatiomlSpace Luunch Strategy study, and the Space Transportation Compankon study for the National Aero-Space Plane Program.

13s= S. Rcpt. 100-326, p. 36, and H. Rept. 100-989, P. 282.

Id’rhe Co$ts md ~nefits of these launch systems would differ; see U.S. Congress, op. cit., foomote 1, and the forthcoming find report of thisassessment.

Chapter 1--Summary ● 5

gram managers to forego these potential savingsand instead pursue greater payload performanceby increasing payload size or weight. If they doso and weight overrun occurs, it would probablycost more to trim the weight of a larger payloadthan it would to reduce the weight of a lighterpayload by theweight margins,reduce this risk.

Option 5:

same percentage. Requiringas described above, would

Congress could continue to fund the develop-ment of the Standard Small Launch Vehicle(SSLV 15) and the Sea Launch and Recovery(SEALAR) system16 to provide survivable meansof launching military lightsats. The Departmentof Defense probably will not allow operationallightsats to be designed for such launch vehiclesuntil the vehicles are operational. Hybrid rock-ets, which can use liquid oxygen to burnnonexplosive solid propellant similar to tirerubber, could also be designed to launchlightsats from transportable launchers. Suchhybrids would have some safety advantages andmight be allowed where conventional solid- orliquid-fuel rockets are not. Later—perhaps by2005-NASP-derived vehicles might be able tolaunch 20,000-pound payloads in wartime.17

With continued funding, the National Aero-Space Plane Program would continue to developtechnology for NASP-derived vehicles.

Option 6:

Congress could fired the development of alaser or a direct-launch system (e.g., a railgun,

coilgun, or ram cannon) for launching micro-spacecraft at high rates economically. Manyuses have been proposed, but to date only theStrategic Defense Initiative Organization (SDIO)has identified a plausible demand for high-ratelaunches of microspacecraft. However, demandfor launches of scientific, commercial, and othermicrospacecraft could increase, perhaps dramat-ically, if launch costs could be reduced to a fewhundred dollars per pound and payloads wereinexpensive. The SDIO estimates developmentand construction of a laser for launching 44-pound payloads would require about $550million over 5 or 6 years. The SDIO estimatesit could launch up to 100 payloads per day (morethan 20 Shuttle loads per year) for about $200per pound.

Railgun proponents predict a prototyperailgun capable of launching 1,100-pound projec-tiles carrying 550 pounds of payload could bedeveloped in about 9 years for between $900million and $6 billion, including $50 million to$5 billion for development of projectiles andtracking technology. If produced and launchedat a rate of 10,000 per year, the projectiles (lesspayload) would cost between $500 and $30,000per pound (estimates differ). The cost of launch-ing them might be as low as $20 per pound—i.e.,$40 per pound of payload.

Is’’f’he SSLV is being develop~ by the Defense Advanced Research projeets Agency (D~A).

IGSEALAR is king developed by the U.S. Naval Research Laboratory.

ITS= U.S. Congress, op. cit., footnote 4, pp. 67 and 74.

Chapter 2

Weight and Volume Margins

As payloads progress from the drawing board (orcomputer-aided design workstation) to the launchpad, their volumes tend to expand to fill the payloadfairings of their betrothed launch vehicles, and theirweights tend to grow to the maximum that the launchvehicles can launch. Dry weights of representativeU.S. military satellites have been about 25 percentgreater, on the average, than the estimates of dryweight made when they were proposed.l On occa-sion, satellites grow so heavy that their assignedlaunch vehicles would be unable to place them in thedesired orbits. When this happens, drastic andexpensive efforts are undertaken to reduce theweight of the satellite—and sometimes of the launchvehicle as well. A TRW executive has said, “Wehave to spend numbers like $150,000 a pound tryingto get the last few pounds out of a spacecraft. ”2

A contract for spacecraft development and pro-duction typically specifies that the spacecraft per-form certain functions and not weigh more than aspecified maximum weight. Usually this maximumweight is chosen to be less than the maximum weighta particular launch vehicle can place in the desiredorbit. The difference between these two valuesprovides a margin that allows for contingencies suchas less-than-expected launch vehicle thrust.

The spacecraft designer initially tries to design thespacecraft to perform the specified functions and notweigh more than the specified maximum weight,minus

1. a “contingency” amount by which the esti-mated weight of the spacecraft is expected togrow as the design matures, and

2. a “weight margin” representing greater-than-expected growth in the estimated weight as thedesign matures.

Based on its experience producing high-tech govern-ment spacecraft and spacecraft subsystems, TRWexpects weight to grow ultimately 15 percent largerthan initial estimates and hence budgets a contin-gency of 15 percent initially. As the design maturesand the estimated weight increases as expected, thecontingency budget is decreased.

TRW may also allow a margin of 15 percent ormore for greater-than-expected weight growth, ini-tially. TRW estimates that the value of weightmargin increases with decreasing margin, from zeroat 15 percent margin, to $5,000 to $10,000 per poundat 5 to 10 percent margin, to $40,000 per pound at Opercent margin, to as much as $100,000 per poundat negative margin, depending on the time remainingbefore launch (figure 2-l). Margins are seldomnegative at an early design stage; if margin becomesnegative shortly before the spacecraft is to belaunched, expensive redesign may be required toreduce the spacecraft weight below the maximumweight specified in the contract, and TRW estimatesthat it maybe worth up to about $100,000 per poundto avoid this. This is much greater than the cost oftransporting a typical spacecraft to its operationalorbit: about $15,000 to $30,000 per pound togeosynchronous orbit, as estimated by TRW.3

TRW notes that spacecraft designed to incor-porate advanced-technology subsystems can con-sume weight margins with unusual rapidity, leadingto redesign (which may require 3 or 4 months),increased risk, and possibly additional redesign, aswell as downward revision of ‘requirements. Two

Figure 2-l—Value of Weight Margin

Marginal value of margin ($1,000 /lb.)100

r - - - ~80 -

II \

60 j ‘,

I-- GEO Iaunch (Titan IV + IUS)

4 0 I

‘“’d --’~-- LEO launch (Titan IV)o

+ — - - = k . — — —— — - — — ~––- — ——T–— = = = + —- 5 0 5 10 15 2 0

Weight margin (percent)

SOURCE: TRW (plotted by OTA).

1P. Hillebrandtet al., space Diviswn Unmanned Space Vehicle Cost Model, Sixth Edition, SD TR-88-97 (Los Angeles Am, CA: Headquarters, SPaceSystems Division, U.S. Air Force Systems Command, Nove~r 1988), p. VIII-6; distribution limited to U.S. Government agencies only.

~.W, Elverum, Jr., “Reliability Up/Coats Down Through Simplicity,” Space Systems Productivity and Man@acturing Conference IV (El Segundo,CA: Aerospace Corp., 1987), pp. 175-1%.

Sclmk ~rby, TRW, prsonal communication, Nov. 16, 1988.

-7–


redesign cycles totaling 6 to 8 months could lead toa 10 to 25 percent cost growth. TRW also notes thatvolume constraints are also costly.

How much margin should be reserved? Reservingtoo little increases the risk of delay and cost overrun,should redesign become necessary, because weightgrowth exceeds expectations late in a developmentprogram. Yet reserving too much margin imposes anopportunity cost: one either foregoes the opportunityto add extra fuel or equipment, and hence capability,to the payload, or one foregoes the opportunity tomake the payload less expensive by allowing it to beheavier. In the latter case, the opportunity cost istangible and can be estimated.4 The optimal marginwill be that at which the marginal opportunity costof not making the payload a pound heavier equalsthe marginal value that the additional margin wouldprovide by reducing risk of cost overrun andschedule slippage.

If the marginal value of weight margin is asestimated by TRW (see figure 2-1), then a margin of11 to 12 percent would be optimal for a Titan IVpayload. 5 This is comparable to the margin (15percent) that Public Law 100-456 required theDepartment of Defense to reserve for satellites DoDapproves for development in fiscal year 1989. Inreporting on the National Defense Authorization Actfor fiscal year 1989, the Senate Committee onArmed Services proposed requiring that

. . . the Under Secretary of Defense for Acquisitionshall not approve for development a new satellite ifthe proposed payload weight exceeds 85 percent ofthe lift capacity of the launch vehicle(s) identified

with the proposed satellite, and shall not approve fordevelopment a block change if the proposed payloadweight exceeds the weight of the existing payload.b

This language was endorsed in conference andsigned into law.8

The expectation that increasing weight marginswould reduce cost risk does not imply that it wouldsave money to build a new launch vehicle largeenough to launch the heaviest payloads with in-creased weight margins. Predicting whether it wouldsave money would require comparing the cost ofdeveloping the vehicle with the total benefits itwould provide—including the reduction in cost riskthat increased weight margins would provide. Ironi-cally, building larger launch vehicles could increasethis risk: Without discipline, payloads might still bedesigned to allow little margin, and margins couldstill be tight or negative, but with greater conse-quence. To eliminate a l-percent negative margin,one would have to trim 2,000 pounds from a200,000-pound payload, compared to only 390pounds in the case of a 39,000-pound payload. Insome cases this could be done by carrying less fuelthan planned and accepting reduced payload per-formance. If not, reducing the dry weight of a200,000-pound payload by 1 percent would cost$240 million more than reducing the dry weight ofa 39,000-pound payload by 1 percent, if weightreduction costs $150,000 per pound. This illustrateswhy allowing margin for, and controlling, weightgrowth will be much more important for largepayloads (such as proposed heavy-lift launch vehi-cles could carry) than for smaller ones.

4sw CCA p~~e~c Analysis” in ch. 3.

Sviz. a paylo~ that would cost between $100 million and $5 billion if built to weigh 39,000 pounds-i.e., to reserve no mwgin.

6s. Rept. 100-326, P. 36.

TH. Rept. I(K)-989, P. 282.8S. ReptO 100-326” did not Swify ~he~er~e [es~ate o~ propo~ Satel]ite weight should inc]ude ex~t~ weight growth in addition tO the IIOIIlllld

weight estimated from the design for the satellite. If so, Public Law 100-456 would require at least a 15 percent weight margin in addition to whateverweight growth is expected. If the payload consisted solely of an unfueled satellite, 25 percent weight growth would be expected [P. Hillebrandt et al.,op. cit., footnote 1, pp. VIII-6], so Public Law 100-456 would require a 40 percent weight margin in addition to the nominal weight estimated from thedesign for the satellite.

Chapter 3

Fatsats

Many experts find it plausible that a payloadcould be designed to perform a function at lower costif it were allowed to be heavier. However, there havebeen few attempts to estimate just how much cheaperpayloads could be, if allowed to be heavier, or howmuch they should weigh in order to minimize thetotal cost of producing and launching a payload.

The first section of this chapter discusses one ofseveral ways in which payloads could be made lessexpensive if allowed to be heavier: standard subsys-tems could be used in lieu of customized subsystemsdesigned to minimize weight. The next sectiondescribes how a high-altitude satellite could beallowed to be heavier (and hence less expensive)without using a larger launch vehicle: by using anorbital transfer vehicle with electric engines (e.g.,arcjets). The third section discusses estimation ofcost versus weight trade-offs, with subsectionsdescribing parametric and bottom-up methods andcomparing parametric with bottom-up estimates.The final section discusses some organizationalobstacles to reducing cost by allowing payloadweight growth.

STANDARDIZING SUBSYSTEMSOne of several ways to trade off cost for weight is

to use standard spacecraft subsystems or busses.2

The use of a standard or previously developedsubsystem may result in a heavier spacecraft butallow a satellite program to avoid paying part or allof the substantial nonrecurring costs of developinga custom subsystem. In addition, because of learningand production-rate effects, it helps reduce therecurring cost of producing the standard subsystem.

Using a standard subsystem could reduce subsys-tem cost by a factor roughly equal to 1 plus the ratio

(expressed as a Ii-action) of nonrecurring to recurringcost. For example, the ratio of nonrecurring cost torecurring cost is typically 2/1 for a spacecraft bus(see table 3-l), so the nonrecurring cost of develop-ing a spacecraft bus is about twice the recurring costof producing a bus. If a mission requires onespacecraft, the cost of developing and producing acustom bus would be about three times the cost ofproducing a suitable previously developed bus. Byusing a previously developed bus, one could there-fore save about two-thirds of the cost of a custom-ized bus.3 The cost would be reduced by a factor of3, i.e., by 1 plus the ratio of nonrecurring cost torecurring cost (2/1).

More could be saved on subsystems with higherratios. For example, the cost of structure, which hasratios ranging from 5/1 to 8/1, could be reduced bya factor of at least 6, and possibly 9.

The amount saved could be a small percentage oftotal spacecraft program costs, which also includethe costs of mission-peculiar payloads and programoverhead, etc. A 1972 Lockheed study4 concludedthat development and use of standard subsystemscould save only about 4 percent of the cost of 91payload programs, when used in addition to low-costdesign methods and payload refurbishment. Thesavings attributable to standardization would beabout 6 percent of the program costs already reducedby low-cost design and refurbishment.5

Manufacturers have estimated that 95 percentlearning might be achieved, i.e., every time thecumulative number of units produced is doubled, theincremental unit cost would decrease 5 percent.6 TheAir Force has assumed 95 percent learning inestimating first-unit production costs from lot sizes

ISW, e.g., LoC~~MiSSileS and Space Co., t~acrOfhW-CoStRefwbiStile and Standard Spacecraft Upon Future NASA S@zce Progmms, NTISN72-27913, Apr. 30, 1972.

2A ~ac=r~’’bm’ consists of hose spacecr~t subsystem~.g. structure, thermal control, telemetry, attitude control, Power, and PmP~sion—thatare not peculiar to a particular mission, as cameras or radio relays would be.

3The cost of ~~~at~g ~ Off..the-shelf subsystem into a spacecr& is sm~l. The Boeing comp~y’s p~~etric Cost Model predicts that the costof integrating an off-the-shelf subsystem into a spacecraft would be about 3 percent of the cost of designing anew subsystem for the spacecraft. [BoeingAerospace Co., May 1989]

o~c~~ Missiles and Space Co., op. cit., footnote 1.5~ his study, ~c~eed ~med a ‘ ‘mission m~el’ now recowi~ w highly i~ated; this probably 1~ to overestimation of the potential savings

from standardization. On the other hand, the percentage savings atrnbutable to standardization might have been greater if refurbishment had not beenassumed.

6F.K. Fong et & um~~ed spacecr@ cost Model, s~ ~., sD.’l’’slAsAs (~ Angeles AF’B, CA: He~qu~ers, space Systems Division, U.S.Air Force Systems Command, June 1981); NTIS accession number AD-B060 824L; disrnbution limited to U.S. Government agencies only.

-9-


Table 3-l-Ratios of Nonrecurring to Recurring Costof Spacecraft Subsystems

If a previously developed subsystem can be used in a new spacecraftin lieu of a custom-designed subsystem, subsystem cost could bereduced by a factor roughly equal to one plus the ratio of nonrecurringto recurring cost-e.g., threefold for a spacecraft bus (a space-craft without its mission payload).

Ranges of ratiosnonrecurring Cost/

Subsystem recurring cost

Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5/1 to 8/1Propulsion (apogee kick) . . . . . . . . . . . . . . . 2/1 to 6/1Thermal control . . . . . . . . . . . . . . . . . . . . . . . 4/1 to 40/1Attitude control . . . . . . . . . . . . . . . . . . . . . . . . 1/1 to 2/1Electrical power . . . . . . . . . . . . . . . . . . . . . . . 1/1 to 2/1Telemetry, tracking, & command . . . . . . . . . 2/1 to 3/1Spacecraft (less mission payload) . . . . . . . . about 2/1Communications mission payload . . . . . . . . 2/1 to 3/1SOURCE: U.S. Air Force Systems Command, Space Systems Division.

and prices,7 but there have been too few buys of eachtype of U.S. spacecraft to demonstrate learningconclusively.

UPPER STAGES AND ORBITAL-TRANSFER VEHICLES

Often a payload has an upper stage or anorbital-transfer vehicle (OTV) in addition to aspacecraft. Some analysts have considered optionsfor reducing the costs of upper stages by allowingthem to be heavier.8 Others are considering theopposite approach: making OTVs smaller-andperhaps more expensive--in order to save money byusing a smaller launch vehicle, allowing the space-craft to be larger, or providing more margin forspacecraft weight growth. Space-based OTVs havealso been proposed;9 they could be reused and wouldnot be launched together with the spacecraft, whichcould therefore be larger. However, refueling andmaintaining them could be complicated and mightrequire the development and maintenance of costlyspace- or ground-based infrastructure.

Electric propulsion could be used to make OTVssmaller and lighter while at the same time increasing.

the mass they could deliver to a high orbit. Theycould use photovoltaic (“solar”) cells to generateelectricity10 to power an electrostatic ion thruster, anarcjet thruster, or an electric engine of some othertype (many are possible) that has an exhaust velocitymuch greater than that of a chemical rocket. Thiswould reduce the mass of fuel required for orbitaltransfer, increasing the payload that could be trans-ferred. The “dry weight” of an electric OTV(EOTV) could also be smaller than that of a chemicalOTV of comparable total impulse, further increasingthe payload that could be transferred.

There would be drawbacks. An EOTV wouldproduce little thrust, so transfer of a payload from alow-altitude parking orbit to geostationary orbitmight take 3 to 6 months. Before reaching itsdestination orbit, the payload would age a fewmonths and, more important, might degrade becauseof its longer transit through the Van Allen radiationbelts. An EOTV would be designed to tolerate sucha transit, but some satellites might not be. Anear-term solution would be to shield sensitivesatellites against the radiation, but this would reducethe maximum satellite mass that could be carried.

The longer transfer time could also be detrimentalto security. A military satellite on an EOTV wouldremain longer at low altitude and within range oflow-altitude anti-satellite weapons than if it rode aconventional OTV to its destination orbit. If acritical satellite fails or is damaged, an EOTV mightnot be able to replace it without a serious lapse inmission performance.ll These drawbacks could bemitigated by launching military satellites on sched-ule rather than on demand. That is, near the end ofits projected useful life, each satellite would bedeactivated, maintained as a spare, and replaced bya new satellite, which could have been launchedmonths earlier.

The Air Force Systems Command Space SystemsDivision estimates that EOTVs would be moreeconomical than conventional OTVs for selected

7p. Hillebrandtet d., space Division unrnan~dspace Vehicle Cost Model, Sixth Edition, SD TR-88-97 (Los Angeles A.FB, CA: Headquarters, SPaceSystems Division, U.S. Air Force Systems Command, November 1988); distribution limited to U.S. Government agencies only.

SD~i ~er, “why Spacwraft Should Get Less Expensive If Launch Costs Decrease, ” unpublished, undated, marked ‘‘ D582-1(X)03-1. ” Boeingpresented results of this analysis at the ALS Phase 1 System Requirements Review, and to OTA in a briefing, “ALS program Development. ” See alsoHughes Aircraft Co., Space and Communications Group, Design Guide for ALS Payloads, October 1988.

gE.g., see ~ckh~d Missiles ~d Space CO., Fiwl Report---payload Eflects Analysis Study, LMSC-A990556 (Sunnyvale, CA: June 1971)! NTISaccession number N71-3749630; and J.M. Sponable and J.P. Penn, ‘‘Electric Propulsion for Orbit Transfer: A Case Study, ’ Journal Of PrOPU/SIOn adPower, vol. 5, No. 4, July-August 1989, pp. 445451,

IONucle~ Wwer was once considered more promising; see Imckheed Missiles and Space CO., Op. Cit., fOOmOte 9.

llHowever, in ~me scenarios, neither could a conventional OTV.

Chapter 3-Fatsats ● 11

missions. For example, a 5,250-pound spacecraftcould be launched to geostationary orbit either on aTitan IV launch vehicle with an Inertial Upper Stagefor about $250 million or on a Delta II launch vehiclewith a solar-electric OTV for about $124 million,12

saving about $126 million. A Titan IV with anEOTV could launch a 30,000-pound spacecraft togeostationary orbit for an estimated cost of $269million. The Space Systems Division is planning todemonstrate an expendable solar-powered EOTV,probably with an experimental payload, sometimebetween 1993 and 1995.

ESTIMATING POTENTIALCOST REDUCTION

There have been few attempts to estimate howmuch cheaper spacecraft could be, if allowed to beheavier, or optimal weights for minimal productionand launch costs. Analyses of historical data showthat heavier spacecraft are typically costlier thanlighter spacecraft;l3 usually, however, heavier space-craft are also more capable than lighter spacecraft.They perform more functions, more difficult func-tions, or similar functions better.14 So these analysesdo not answer the important questions:

1.

2.

How heavy should a payload15 be in order tominimize the combined costs of payloadproduction and launch on a currently opera-tional launch vehicle? Are payload weightsactually optimized for current launch vehicles?How heavy should a payload be in order tominimize ‘combined production and launchcost, if the launch cost per pound of payloadwere reduced, or the maximum payload weightthat could be launched were increased, bysome factor? By what factor would totalpayload production and launch cost be re-duced?

Answering these questions requires comparingthe costs of heavy payloads and light payloads thatperform the same function equally well. Unfortu-nately, such data do not exist; there are no two

payloads, one large and the other small, designed atthe same time (and hence with comparable technol-ogy available) to perform the same set of functionsequally well. The few analyses that have estimatedhow much cheaper a payload could be if allowed tobe heavier have been hypothetical. They are basedon both “bottom-up” and parametric estimates.

Bottom-up estimates are obtained by designingtwo or more versions of a payload to perform thesame functions at minimum cost without exceedingweight or size limits, which differ from version toversion. For example, two versions of a communica-tions satellite could be designed: one to be launchedon a Scout launch vehicle, the other on an Atlas-Centaur. Each version would be designed to mini-mize production and launch cost. Comparing thecosts of the two different versions would indicatehow much less expensive the larger version wouldbe.

Bottom-up estimates are time-consuming andexpensive to derive, and there may be no basis forassuming the cost-versus-weight trade-offs derivedwould apply to versions larger or smaller than thosedesigned or to payloads that must perform differentfunctions. For example, there is no rationale forexpecting that bottom-up estimates of costs andweights of communications satellites of comparablecapability could be used to illustrate cost-versus-weight trade-offs for remote-sensing satellites ofcomparable capability.

Parametric estimates are obtained by assumingthat if the weight of a payload were allowed toincrease, the minimum cost at which it could be builtwould vary in some qualitative way-e. g., approacha limit, or decrease exponentially. The parameters ofthe relationship-e.g., the minimum costs for partic-ular weight limits—are chosen to make the hypo-thetical relationship fit historical cost and weightdata, bottom-up cost and weight estimates, or both,as well as possible.

The fit will not be exact, however;l6 it may be agood fit on the average, with some payloads costing— - - —

lz~cluding $3 million for RDT&E.

13P. Hillebrandt et al., op. cit., footnote 7, and figure 4-1 of Lockheed Missiles and Space CO., Op. cit., foomote 9.

14s= ~ckh~ Missiles and Space Co., op. cit., footnote 9.15Thepq/o~ of a lawch vehicle may ~ a ~acar~t (e.g., a satellite or plane~ probe) toge~er wi~ an upper stage (to propel it to a transfer orbit

or escape trajectory) and support equipment for attaching them to and releasing them from the launch vehicle, Some launch vehicles (e.g., the SpaceShuttle and sounding rockets) sometimes carry payloads (e.g., scientific instruments) that remain attached to the launch vehicle.

16u~e= mere ~ s. many Pwmeters (i.e., statistic~ degrees of fi~om) in tie model mat tie av~lable data ~e too few to estimate them with

statistical significance.


more, and others less, than predicted by the relation-ship (the “model”). The fact that the modelrepresents only average costs and cost-versus-weight trends may be an advantage in studies, suchas this one, which focus on such averages and trends.However, it would be a limitation if accuratecost-versus-weight trade-offs for a specific missionwere required. If the functions to be performed bythe payload were specified in detail, and if resourcespermitted detailed engineering design of severalalternative versions of different weights or sizes,then bottom-up estimation could be used and shouldprovide greater accuracy.

Thus bottom-up estimation and parametric esti-mation are complementary approaches to cost esti-mation.17 Neither approach by itself would be ofgeneral value: a bottom-up estimate is applicableonly to a specific payload, and parametric estimatesare abstract and hence useless unless fitted tobottom-up estimates of cost and weight.18

The bottom-up and parametric approaches areexemplified by two analyses produced almost twodecades ago: a parametric model developed by CarlBuilder at the Rand Corp.,19 and a bottom-upanalysis by Lockheed Missiles and Space Co.20 Bothwere cited in congressional debate on the merits ofthe Space Shuttle.21

A Parametric Analysis

Builder did not estimate cost reductions forparticular payloads; he described a procedure fordoing so using data or assumptions about payloadcost and launch cost. He assumed average launchcost would vary as the payload capability of thelaunch vehicle raised to some power A (this is calleda “power-law” relationship) and payload costwould vary as the payload weight raised to some

power B. To use Builder’s model, one must specifythe exponent A and the initial cost and weight of apayload designed to minimize payload plus currentlaunch cost. One need not specify what the payloaddoes; in theory, it should make no difference. Iflaunch costs are reduced by some factor,22 theoptimum weight and the cost of a functionallyequivalent payload designed to minimize payloadplus reduced launch cost may be calculated usingformulas derived by Builder. The difference be-tween the old and new minimum total costs is thesavings obtainable if launch costs are reduced by thespecified factor, assuming payloads are reoptimizedto take advantage of the new launch costs.23

OTA derived an estimate of the exponent A usedin Builder’s model by fitting a straight line to pointson a log-log plot of the payload capabilities24 andaverage launch costs25 of Delta, Titan, and the SpaceShuttle. Figure 3-1 shows the three points and theline obtained as a least-squares fit to the points. Theslope of the line corresponds to a value of 0.74 forthe exponent A, implying that average launch costwould increase by two-thirds if the payload weightwere doubled.

Figure 3-1 also shows a point representing thepredicted payload capability of the Pegasus air-launched vehicle (see figure 4-6) and the pricecharged by its operator, Orbital Sciences Corp.(OSC), for the launches and launch options pur-chased by the Defense Advanced Research ProjectsAgency. The line fitted to Delta, Titan, and Shuttlecosts and payload capabilities accurately predictedthe cost (to the government) of a Pegasus launch.

If launch cost is assumed to vary with payloadweight in the way described above,26 Builder’smodel predicts that a payload that would cost $1billion if designed to weigh 39,000 pounds could be

17u.s. @gess, office of Technolo~ Assessment, Reducing Launch Operations Costs: New Technologies and Practices, OTA-TM-ISC-*8(Washington, DC: U.S. Government Printing Office, September 1988), app. A.

18As no~ a~ve, hem is seldom more than one historical “data point” for the cost and weight of fictionally identical SPaceraft.19cwl H. BUil&, Are ~~h Vehicfe c’05f5 a Bottle~~k t. ECO~rniCa/spuce operatio~.? Rand Working document D-19482-PR, December 1969.

%ckheed Missiles and Space Co., op. cit., footnote 9.21s= ~erem~s of Senatm Mond~e in tie CongresswMl Recor#e~te, May 26, 1971, pp. 171~-]7106, and tie rem~ks of Senator Anderson

in the Congressional Record-Senate, Apr. 20, 1972, pp. 13786-13790.22BY tie me factor, regardless of payload weight.

~More P=isely, “mw~g paylo~ we optim~]y si~ for minimum total cost in both cX.WS.” Builder, OP. cit.! foomote 19.

24T0 a ~oonemi.-hi~ Ofiit inc~ined 28.5 de~s. U.S. Congess, Office of T~hno]ogy As~ssment, ~nch Optzonsfor the Future: A &4yer’s Gw”de,OTA-ISC-383 (Washington, DC: U.S. Government Printing Office, July 1988), table 2-1.

2S~cluding ~m~ fix~ amu~ ]aMch co~t, OTA ~~~ launch at tie m~imum rate es~at~ in table A-1 of ibid. and usd the nominalcost-estimating relationships in that table.

~Ioee, in ~oP~m t. tie paylo~ capability of the launch vehicle raked to tie Power 0-74.

Chapter 3—Fatsats ● 13

Figure 3-l—Average Launch Cost v.Payload Capability

Average launch cost (millions of dollars)$l,ooo ~– ---- -- -

~$100 ~

, , p - ‘

~

I, , . F ’ “

$10- , -

1●-

+ , , ”$ 1 ‘ “ : ’ - ’ “- - — — T – — — ” — - – - — - - -

0.1 10 100Payload’ capability (1,000s of Ibs.)

Power-law fit ● Delta + Titan

* S h u t t l e ■ Pegasus (not fitted)

SOURCE: Office of Technology Assessment and Orbital Sciences Corp.,1988.

made for 5.8 percent less if it were allowed to betwice as heavy. More generally, if allowed to beheavier or designed to be lighter, the cost of such apayload would be proportional to its weight raised tothe power –0.086 (the exponent B in Builder’smodel). Figure 3-2 shows how the payload cost andthe launch cost would vary with the payload weight.Designing the payload to weigh 39,000 poundswould minimize the total cost.

Figure 3-3 shows how the payload cost and thelaunch cost would vary with the payload weight iflaunch cost were reduced by a factor of 3—i.e., by 67percent. It illustrates that total cost is insensitive toweight for weights between 80,000 pounds and (atleast) 200,000 pounds. The optimal weight would beabout 150,000 pounds. Reducing the launch cost by67 percent would reduce the total cost by only 11percent. It should be emphasized that this is theestimated cost reduction achievable by allowingpayload weight to grow without changing payloadperformance. It assumes that the baseline payloadwas designed to minimize total cost. If a baselinepayload was not designed to minimize total cost,redesigning it (possibly to weigh more) could savemoney even if launch costs are not reduced. If launchcosts are reduced, additional savings could beobtained by allowing weight growth; these addi-tional savings are the savings estimated by Builder’smodel.

Figure 3-4 shows these results in a different formalong with results of similar analyses of otherhypothetical payloads initially weighing 39,000pounds but costing $2 billion, $3 billion, and $4

Figure 3-2-Cost for Hypothetical Mission WithCurrent Launch Cost Trend

Cost (millions of dollars)$1,400 ;–

$1,200 “

$1,000-,

$8001

$ 6 0 0

$400$200 ~

$0 , - — — —10 30 50 70 90 110 130 150 170 190

Payload weight (1,000s of Ibs.)

r P a y l o a d m L a u n c h- —-

SOURCE: Office of Technology Assessment, 1990.

Figure 3-3-Cost for Hypothetical Mission if LaunchCosts Are Reduced 67 Percent

Cost (millions of dollars)$1,400 ~-- - - — – - 1$1,200

$1,000$800 -

$600 -

$400 -

$200 -$ 0 - - — ~

10 30 50 70 90 110 130 150 170 190Payload weight (1,000s of Ibs.)

~ P a y l o a d _ L a u n c h


billion. Figure 3-4a shows how these costs could bereduced, according to Builder’s model, if theweights were allowed to grow. For each weight,figure 3-4b shows the (“economic’ launch cost atwhich that weight would be optimal. These esti-mates predict that allowing the weight of a Titan-IV-class payload to increase by 400 percent (forexample) would reduce payload cost by an amountthat is nearly the same for a $1-billion payload as fora $4-billion payload. The estimates also predict thatthe lower the initial cost of a payload, the more thecost per launch must be reduced to justify increasingits weight by a large factor.

Builder’s assumption about the relationship be-tween payload weight and launch cost would lose

. .


Figure 34-Payload Cost v. Weight Trade-offs and—

Payload$5

$4

$3

~$2

$1-

$0 ~. ——

0

Economic Launch Costs

cost (billions of dollars) A— —. — —

— — . . — . I

- – - — — . . 7 — — . — — —50 100 150 200

Payload weight (1,000s of Ibs.)

Economic launch cost (millions of dollars) B$ 1 2 0 ,

$115-

$110-

-

$105- - d

+ –

$ 1 0 0 L —0

‘ \ - — – - . .+P a y l o a d 1 ‘ \ ~

Payload 2

Payload 3

‘ .l ,

= . -.Payload 4 - = . , ----- -.. .

. —= - — — - r – — ~

50 100 150 200LV payload capability (1,000s of Ibs.)


validity if extrapolated to extremely heavy pay-loads.27 It would imply that the average cost oflaunching a pound of payload could be made as lowas desired--even lower than the cost of the fuelrequired to launch a pound of payload—by buildinga launch vehicle of sufficiently large payloadcapability. But his assumption fits the estimates ofPegasus, Delta, Titan, and Shuttle launch costs in

figure 3-1 very well. The cost-versus-payload curvefitted to Delta, Titan, and Shuttle launch costspredicted the cost of Pegasus accurately, eventhough the payload capability of Pegasus is eightfoldsmaller than that of the smallest vehicle (Delta) onwhich the curve is based. The curve could probablybe extrapolated with comparable validity to apayload capability of 200,000 pounds.

Similarly, Builder’s assumption about the rela-tionship between payload weight and payload costwould also lose validity if extrapolated to extremelyheavy payloads.28 It would imply that they could bebuilt at a lower cost per pound than that of the bulkstructural material (e.g., aluminum) from which theyare made. However, this is not a problem for theranges of payload weights and costs-shown in figure3-4.

Bottom-Up Analyses

Lockheed used bottom-up analysis to estimatehow much the cost of building, launching, andoperating selected payloads could be reduced bymaking them larger29 and by other measures.30 Lock-heed considered three payloads, selected to span arange of costs, that had been built and launched andfor which design and cost data were available. Theleast expensive was the Lockheed P-1 1 subsatellite,which could be modified for use as a Small ResearchSatellite (SRS). The most expensive was the Orbit-ing Astronomical Observatory, the redesigned ver-sion of which was designated OAO-B. The otherwas the Lunar Orbiter, which could be modified foruse as a Synchronous Equatorial Orbiter (SEO), fourof which could perform Earth resources observation.

Lockheed estimated how much the costs31 of theOAO-B and SEO payload programs could bereduced if the payloads were redesigned to belaunched on unmanned expendable launch vehiclesor to be launched on a (then) proposed version of theSpace Shuttle.32 The savings estimated for the first

A is less than one.

B is less than one. t. for

trade-off curve for each subsystem was a which, for a minimum cost per pound and, at the otherextreme, approached the minimum weight achievable.

refinishing in orbit or retrieving to repaired or on development, testing, if the Shuttle

used, refurbishment) of satellites. considered two versions.


case were attributed to payload growth.33 However,Lockheed did not redesign the baseline versions ofOAO-B and SEO to minimize cost without exceed-ing the baseline weight, hence Lockheed attributedto weight growth some cost reduction that shouldhave been attributed to improved design. Theamount of cost reduction misattributed to weightgrowth cannot be determined from Lockheed’sreport, so the cost reductions Lockheed attributed toweight growth should be considered upper boundson cost reductions achievable by allowing weightgrowth.

Figure 3-5 compares Lockheed’s estimates of theweights and the average unit costs34 of the baselineOAO-B and SEO with Lockheed’s estimates of theweights and costs of “low-cost” versions designedto be launched on expendable launch vehicles. Thepotential savings in fiscal year 1988 dollars wouldbe $10.1 million (21.3 percent) and $43 million(15.2 percent) for SEO and OAO-B, respectively.The estimated weight growth required to achievesuch savings would be 170 percent and 69 percent,respectively.

A more recent bottom-up analysis by BoeingAerospace Co. estimated the cost of a “typical”payload could be reduced by a large percentage ifweight growth by a modest percentage were al-lowed, and that it would save money to allow suchweight growth if launch cost per pound werereduced.35 For example, Boeing estimated the costcould be halved if the weight were allowed to grow30 percent (see figure 3-6).36

Boeing actually considered a payload consistingof an upper stage and a hypothetical spacecraft usingspecific subsystem technologies and with subsystemweights in an assumed ratio.37 Boeing claimed thehypothetical spacecraft was typical, implying thatsimilar cost reductions could be expected for othertypes of spacecraft. However, some of the subsystemtechnologies Boeing assumed for the spacecraftwere atypical. For example, the analysis estimated

Figure 3-5-if a Satellite Were Allowed To Be Heavier,It Could Cost Less

Average cost (FY 1988, millions of dollars)1 , 0 0 0 ‘- —

I

—

t Baseline

1 * — “Low-cost ”

1

Orbiting *

1 0 0Astronomical

Baseline Observatory-B. .Low-Cost ”

➤ .

I Synchronous

t EquatorialOrbiter

1

10 ‘——-- ‘-”– , . - - ~1,000 10,000

Weight (Ibs.)

SOURCE: Estimates from Lockheed Missiles and Space Co., 1971, plottedby Office of Technology Assessment, 1990.

Figure 3-6-The Effect of Reducing Launch Cost onthe Optimal Weight and Cost of a Payload

Optimal growth~ _ — –

1

factor

‘ — - - .l

- - t - + ”

0“ - — — — - 1

0 1,000 2,000 3,000Launch cost (dollars per pound)

4,000

Boeing Aerospace Co. estimated that, if average launch cost perpound were reduced from $3,600 per pound, it would beeconomical to redesign a hypothetical payload to allow it to beheavier by the weight-growth factor indicated and less costly bythe cost-reduction factor indicated. Boeing assumed that theoriginal design minimized the sum of payload and launch costs ata launch cost of $3,600 per pound, and that the average launchcost per pound does not depend on the payload weight.SOURCE: Boeing Aerospace Co., 1988.

savings estimated for attributed to payload growth; the rest was attributed to reduced cost and to that only a reusable vehicle such the Shuttle could provide: intact abort capability, on-orbit checkout, repair, and refurbishment. Additional savingsin both cases were attributed to use of improved technology.

number of satellites launched (6 to

by OTA. Op. cit., footnote 8.

have misquoted would

growth. See, e.g., Thomas M. ‘Status of the ALS Program,’ Space Systems Productivity and Conference-V (El CA: TheAerospace Corp., 1988), p. 30.

to change when the payload was redesigned with relaxed weight


the cost and weight of the satellite’s electric powersubsystem by assuming it consisted entirely of solarcells (which typically dominate the cost of asatellite’s power subsystem), with no batteries(which typically dominate the weight of a satellite’spower subsystem).

Some of the proposed cost-reducing and weight-increasing substitutions Boeing proposed--e. g., sub-stitution of commercial-grade solar cells for spacecraft-grade (S-class) solar cells—would decrease payloadreliability and expected lifetime to an extent notestimated by Boeing. Thus Boeing did not estimateweight-versus-cost trade-offs for equal reliability;some of the savings Boeing attributed to weightgrowth should have been attributed to reducedreliability .38 Therefore, the savings Boeing attrib-uted to weight growth may be upper bounds forspacecraft of the type Boeing considered.

Another recent bottom-up analysis estimatedcost-versus-weight trade-offs for some subsystemsbut not for complete payloads.39 Like earlier studiesby Lockheed40 and the Aerospace Corp.,41 it identi-fied payload “cost drivers"—i.e., costly payloadcomponents or testing-and recommended changesin payload design, components, testing, or opera-tions that might reduce space program cost. Many ofthese changes would require increasing the weight ofthe payload (the “fatsat” approach discussed here)or specifying a simpler or easier mission, or fewermissions (the “lightsat" approach discussed inchapter 4).

A Comparison of Parametric andBottom-Up Analyses

Estimates differ on how much cheaper payloadscould be if they were allowed to grow to a specifiedweight, and how much they should grow if cost perlaunch were reduced to a specified amount. Oneparametric estimate by OTA predicts that a hypo-thetical expensive payload as heavy as a Titan IV

could launch could cost about $130 million less ifallowed to be five times as heavy. It would beeconomical to design the payload to be so heavy ifit could be launched for less than about $100million. 42 Less would be saved if the baselinepayload cost were comparable to or less than theaverage Titan IV launch cost, estimated hereas$117million. The only bottom-up estimate that could becompared to these is one by Boeing, which predictsmuch greater savings (at least 76 percent, or $760million for a billion-dollar payload) but is based ona conceptual design for a payload that is atypical inimportant respects; moreover, the redesigned pay-load was allowed to use less reliable components,and launch cost per pound was assumed to beindependent of payload weight.

OTA also derived parametric estimates to com-pare to detailed bottom-up estimates by Lockheedfor two spacecraft. Lockheed estimated 66 percentgreater savings for one spacecraft (SEO), and 360percent greater savings for the other (OAO-B), butattributed to weight growth some savings that shouldhave been attributed to optimization of the baselinedesigns. However, these discrepancies are compara-ble to the unexplained statistical variations oftenencountered in spacecraft cost estimation. The lessdetailed Boeing estimate, if applicable, would pre-dict much greater savings than predicted by OTA:530 percent more for OAO-B, and at least 630percent more for SEO, at the weight growth factorproposed by Lockheed.43

ACHIEVING POTENTIAL COSTREDUCTION

Realizing most of the potential savings predictedby these estimates will probably require creation ofincentives to dissuade satellite program managersand designers from adding capability, and therebyweight, until launch vehicle lift margin, engine-out

sg~ckh~dids. and estimated how many satellites would be required to provide comparable mission performance for 10 years. bckh~d concludedthat use of S-class components, which would maximize expected satellite life and minimize the number of replacements required, would be mostcost-effective.

39Hughes Aircraft Co., op. cit., footnote 8.

%ckheed Missiles and Space Co., op. cit., footnote 9.blspacecra~ Cost Drivers Study--+inal Report: Phase 1 (El Segundo, CA: The Aerospace COW., October 1983).

bz’’f’he Advanced Launch System is intended to launch such payloads fO1 much less.b3However, Boe~g estimated that it would not be economical to ~ek the ex~eme cost reduction predicted for SEO unless launch cost per Pourtd were

reduced by a factor of 36.


Box 3-A—Fatsats: The Good and Ugly, and the BadHEAO—During a budget crunch, NASA took steps to repackage its High-Energy Astronomical Observatory

instruments, designed for two Titan payloads, into three Atlas-Centaur payloads. This reduced launch costs by about2 x $125K -3 x $60K = $70K in today’s dollars, by TRW’s calculation. It also created extra weight margin. Thisallowed designers to design the spacecraft with high safety factors (strength margin etc.), so that they could dispensewith the costly construction and testing of model and qualification spacecraft.l

Phobos 1 & 2-In July 1989, two science spacecraft, Phobos 1 and Phobos 2, were launched from the SovietUnion toward Phobos, the larger of the two moons of Mars. Their busses were designed and built in the SovietUnion, as were some of their instruments; other instruments were designed in Austria, France, Sweden, Switzerland,West Germany, and several East European nations participating in Project Phobos. Some of the Soviet instrumentswere designed with generous weight margins. Jochen Kissel, a West German member of the project’s scientificcouncil, said, “We could use standard printed circuit boards rather than ultraminiaturized parts . . . It madeeverything cheaper and simpler.

Nevertheless, because of greater-than-expected weight growth, some instruments were removed from one orthe other spacecraft. The two spacecraft were originally intended to carry identical suites of instruments, so thatPhobos 2 could perform all functions of Phobos 1 in case Phobos 1 failed, or vice versa.

As it happened, Phobos 1 did fail. More accurately, contact with Phobos 1 was lost late in 1988, because anerroneous command was transmitted to Phobos 1 from the ground. To compensate for the loss, mission directorsplanned to command Phobos 2 to rendezvous with Phobos rather than proceeding to the smaller Martian moon,Deimos, as it would have had Phobos 1 succeeded. Phobos 2 lacks the radar mapper, neutron spectrometer, and solarx-ray and ultraviolet telescopes of Phobos 1, but carries an infrared spectrometer and a hopping lander which Phobos1 lacks.3

Ironically, contact with Phobos 2 was also lost on March 27, 1989, about two weeks before the plannedencounter with Phobos was to occur (on April 9-10).4 Nevertheless, Phobos 2 gathered a significant amount of databefore this failure. This anecdote illustrates the value of generous weight margins on some instruments, the cost ofnegative weight margins on other instruments and on the spacecraft as a whole, the value of redundant spacecraft,and the risks of human errors and compound failures.

Milstar-Milstar is an advanced communications satellite being built for the Department of Defense. A fewwill be built—fewer than originally planned—allowing nonrecurring program costs to be amortized over a fewsatellites. Aside from this economy, Milstar appears to exemplify the antithesis of the fatsat philosophy: it isdesigned to be large in order to cram it with capability, not to reduce its cost. It has had to be redesigned at leastonce to reduce its estimated weight and add margin. Costly edge-of-the-art technologies have been adopted to reduceweight, and some are so risky that additional greater-than-expected cost and weight growth could occur.

Milstar was made fatter to be better, not cheaper, and its gross weight kept growing. A subsystem designerquipped, “Milstar is going to gross everybody out. ”

ITRW, briefig, NOV. 16, 1989. See also Science, vol. 199, Feb. 24, 1978, p. 869, and Astrophysical/Journal, vol. 230, June 1, 1%’9, P. 540.

2EI-iC J. hrner, “Mission to Phobos,” Aerospace America, September 1988, pp. 34-39.

31bid., and “Phobos 1 Loss to Change Mars MiSSiOn,” Aviation Week and Space Technology, Oct. 3, 1988, p. 29.4C6Soviets IJJSe Contact With Phobos 2 Spacecraft,” Aviation Week and Space Technology, Apr. 3, 1989, p. 22.

capability, and reliability are reduced, requiring spend less than the amount appropriated. Whenexpensive weight reduction programs for redress. funding allows, spacecraft purchasers, with rareThe managers of satellite programs funded by exceptions, opt to increase performance rather thanline-item appropriations have little incentive to reduce cost.

Chapter 4

Lightsats

Lightsats are satellites that are light enough to belaunched by small launch vehicles such as a Scout orPegasus, or others now in development. Militarylightsats could be designed for wartime deploymentor replenishment from survivable transportable launch-ers to support theater commanders. Civil lightsats,and some military ones, would not require transport-able launchers; they could be launched by a widervariety of launch vehicles, including larger ones.

The first Soviet and U.S. satellites were lightsats,according to this definition. Explorer I, the first U.S.satellite, weighed only 31 lb but collected data thatled to the discovery of the Van Allen radiation belt.But in another sense, the first satellites were fat—asheavy as early launch vehicles could launch. Aslarger launch vehicles were developed, larger, morecapable satellites were developed to ride them.Nevertheless, small satellites continue to belaunched for civil and military applications thatrequire only simple functions.

Interest in lightsats has grown recently, l partly inanticipation of new rockets designed to launch themat low cost, and, in the case of military lightsats,because of a desire for a survivable means oflaunching satellites-e.g., transportable launch ve-hicles too small to launch large satellites.2

A few years ago the Defense Advanced ResearchProjects Agency (DARPA) began to examinelightsats, initially to demonstrate the ability ofsimple and inexpensive satellites to perform simplebut useful tasks, and, more recently, to demonstratethe utility of satellites small enough to be launchedfrom transportable launchers to support theatercommanders during a war. DARPA is consideringconcepts for several types of lightsats--for communica-ions, navigation, radar mapping, and targeting. TheArmy, Navy, or Air Force may choose to procuresimilar lightsats for operational use. They would bedesigned to be affordable as well as small, becausemany might be needed for replenishment afterattrition. “With lightsat, we can undoubtedly put upsatellites for less money than it will cost the Soviets

to shoot them down,’ said Dr. John Mansfield,while Director of DARPA’s Aerospace and Strate-gic Technology Office.3

SPACECRAFT CONCEPTSThe first lightsat developed by DARPA’s Ad-

vanced Satellite Technology Program was a commu-nications satellite, the Global Low-Orbit MessageRelay (GLOMR; see figure 4-l). Weighing only 150lb, GLOMR was launched by the Space Shuttle onOctober 31, 1985, into a 200-mile-high orbit in-clined 57 by degrees.

DARPA has ordered nine more UHF communica-tion satellites from Defense Systems, Inc., thecontractor that built the GLOMR. Seven of these,called Microsats, will weigh approximately 50 lbeach. These satellites will be launched together intoa polar orbit 400 nautical miles high by the Pegasuslaunch vehicle. Once deployed, they will spread outaround the orbit. They will carry ‘‘bent-pipe’ radiorepeaters—i.e., the messages they receive will beretransmitted instantaneously to ground stations.The other two satellites will be larger “store/foreward" satellites called MACSATs (forMultiple-Access Communication Satellite). They will weighabout 150 lb each and will be launched together ona Scout launch vehicle. They will store messagesreceived from ground stations and forward (or“dump”) them when within range of the groundstations to which the messages are addressed. Thebent-pipe satellites and the first store/forewardsatellites will cost DARPA about $8 million exclud-ing launch costs.

Amateur (“ham”) radio operators have built aseries of small satellites carrying radio beacons orrepeaters, the first of which, OSCAR I, was launchedin 1961. Since 1969, the nonprofit Radio AmateurSatellite Corp. (AMSAT) and its sister organizationsworldwide 4 have built or designed several of thesefor scientific, educational, humanitarian, and recrea-tional use by hams. AMSAT, which has only onepaid employee, is now building four 22-pound

IS=, e.g., A.E. Fuhs and M.R. Mosier. ‘‘A Niche for Lightweight Satellites,’ Aerospace America, April 1988, pp. 14-16., and Theresa M. Foley,“U.S. Will Increase Light.sat Launch Rate to Demonstrate Military, Scientific Uses, ’ Aviation Week and Space Technology, Sept. 26, 1988, pp. 19-20.

z~ere is no consensus on whether this is the best approach to assuring continued mission performance in w~ime.

@ot~ by James W. Rawles, “LIGHTSAT: All Systems Are Go,” Defense Electronics, vol. 20, No. 5, May 1988, p. 64 ff.4AMSAT-Nofi America hm sister organizations in &gentina, Aus~alia, Br~il, Britain, @-many, Italy, Japan, Mexico, and the Netherlands. The

Soviet Union has also launched satellites for amateur radio operators.

–19–


Figure 4-l-The Global Low-Orbit Message Relay(GLOMR) Satellite

SOURCE: Department of Defense.

(l0-kilogram) communications satellites called Mi-crosats5 (figure 4-2). (These Microsats are unrelatedto the above-mentioned Microsats developed forDARPA and to the “microspacecraft” discussed inthe next part of this report.) All four satellites aredesigned to receive, store, and forward digitalmessages using a technique called packet communi-cations. All four use a standard bus, the Microsatbus. One of the satellites, called PACSAT, is beingbuilt for AMSAT. An almost identical satellitecalled LUSAT is being built for a sister organization,AMSAT-LU, in Argentina. A third satellite, nick-

named Webersat, will carry, in addition to its packetradio repeater, a low-resolution color TV cameradesigned at the Center for Aerospace Technology atWeber State College in Logan, UT. The fourthsatellite, Digital Orbiting Voice Encoder (DOVE),was built for BRAMSAT (AMSAT-Brazil). It willcarry a digital voice synthesizer to generate voicemessages that can be received by students usinginexpensive“scanner’ radios.6

AMSAT contracted with Arianespace to launchthese four satellites together for $100,000. TheAriane 4 launch vehicle will also deploy theUoSAT-D and UoSAT-E amateur-radio satellitesbuilt at the University of Surrey in England inaddition to the four Microsats and its primarypayload, the SPOT 2 photomapping satellite.7 Thelaunch, originally scheduled for June 1989, has beenpostponed until January 1990, at the earliest.

Microsats are among the smallest communica-tions satellites ever built. They are lightsats, becausethey are designed to perform relatively simplefunctions. But they are also fatsats, because they areheavier and larger than they would be if built bymethods usually used for more conventional (andmore expensive) satellites. Assembly of some Mi-crosat subsystems is literally a cottage industry.Although some printed circuits are being built forMicrosats by a contractor using high-tech methods,others are being built in the homes of Amateur radiooperators all over the country. When they volunteerto assemble printed circuits, AMSAT sends them theinstructions.

CONCEPTS FOR PHASEDARRAYS OF SPACECRAFT

Groups of small satellites could collectivelyprovide communications or radar capabilities thatcould otherwise be provided only by a large satellite.To do so they would have to operate coherently aselements of a phased array: all satellites must relaythe signals they receive to a satellite or groundstation that can combine them in a way that dependson the relative positions of the satellites, which mustbe measured extremely accurately. When transmit-ting, the satellites must all transmit the same signal,

“The AMSAT-NA 73 Amateur Radio, May 1989, 83-84, and Doug and Bob The Next Generation of OSCAR Satellites—Part l,” May 1989, pp. 37-40.

digital voice synthesizer is similar to the developed at the University of Surrey in England for use on the and OSCAR-11 amateur-radio satellites built there.

Ward, “Experimental OSCARS,” May P. 62

Chapter 4--Lightsats ● 21

Figure 4-2--Microsat: A 22-lb Communications

(1) Flight model beside Leonid Labutin, UA3CR.(r) Close-up view. (Copyright Radio Amateur Satellite Corp.)Photos: Andrew C. MacAIlister, WA5ZIB.

but each satellite must delay its transmission by aperiod that depends on its relative position.

The Air Force has considered “placing largephased arrays in space with major components of thearrays not rigidly connected to each other” (seefigure 4-3), because “If we can achieve coherenceamong these components, phased arrays can bespread out over very large volumes in space, givingthem an unprecedented degree of survivability. Ittherefore may be possible to create a phased-arraydevice (e.g., a space-based radar) that we can placeinto space and enhance simply by adding morerelatively inexpensive elements whenever the threatincreases and budget pressures permit.”8 If smallenough, each element could be launched by a smalllaunch vehicle.

If a large radar or communications satellite weredivided into several modules, “crosslink” equip-ment for communications among the modules wouldhave to be provided, which would add some weightand cost. Aside from this, historical cost data

indicate that a communications mission payloadmight cost less if divided into several smallerpayloads of equal aggregate weight and power.9 Thesame might be true of radar equipment. Each modulewould need its own bus, but the cost data alsoindicate that several small busses would probablycost less, or no more, than a single bus of equalaggregate weight.10 Learning and production-rateeffects could make the small modules and busseseven less expensive. Economies of scale in launch-ing might make it economical to launch as many aspossible on a large launch vehicle, but they could belaunched individually on small launch vehicles ifdesired-e. g., in wartime, if peacetime launch facili-ties have been damaged.

Coherent operation of several satellites requiresrelative positions to be measured with errors nogreater than a fraction of a wavelength of theradiation to be sensed or transmitted. The accuracyrequired for coherent operation of several satellitesas a microwave radar or radiotelescope has alreadybeen demonstrated.11 Coherent operation of several

Command, Project Forecast Executive Air Force Base, MD: Headquarters, Air ForceSystems Command, undated). This concept is described in greater detail in the classified report. see PP. to of Annex D of vol. IV, which authorized readers may request from the Defense Technical Information Center (accession number 642).

Division Edition, Angeles AFB, CA: Headquarters, Systems Division, U.S. Air Force Systems Command, November 1988); distribution limited to U.S. Government agencies only.

Ground Antennas Link for Radio Astronomy Observations, ”Aviation Week and Space Technology, Dec. 1, 1986, pp.32-33.


SOURCE: U.S. Air Force.

satellites as a ladar or optical telescope is beyond thestate of the art. However, someday it may be feasibleto launch several telescope modules, each smallerthan the Hubble Space Telescope, and assemblethem in space (or allow them to assemble them-selves) into a rigidly connected phased array12 thatwould operate as an optical telescope with betterlight-collecting capability and resolution than theHubble Space Telescope. If technology advancesfurther, two or more such arrays, not rigidly con-nected to one another, could operate coherently tofurther increase light-collecting capability and, es-pecially, resolution.13

A scientist at the Jet Propulsion Laboratory (JPL)of the California Institute of Technology (CalTech)

has proposed a less ambitious phased-may radio-telescope that could be begun today: an OrbitingLow-Frequency Array of 6 or 7 satellites in aformation 200 km across (see figure 4-4). It couldmap astronomical sources of radio signals withwavelengths longer than 15 meters; such signalscannot penetrate the Earth’s ionosphere to reachground-based radiotelescopes. The angular resolu-tion of the proposed array could be comparable tothat of a dish antenna 200 km across. JPL estimatesthat each satellite would cost about $1.5 million andweigh less than 90 kg (200 lb) if cylindrical, or 45 kg(100 lb) if spherical14 They must be launched to acircular orbit at least 10,000 km high, so theequivalent weight to low orbit would be about 170lb for the spherical satellite.

LAUNCH REQUIREMENTSOperational launchers for military lightsats must

meet several requirements, the most distinctivebeing survivability in high-intensity (if not nuclear)conflict. Such survivability is required of U.S.strategic and theater missile launchers, and opera-tional lightsat launchers might adopt some of theirfeatures. Like the rail-mobile launcher for the SovietSS-24 intercontinental ballistic missile (ICBM), thesimilar launcher being developed for U.S. Peacekeeper(M-X) ICBMs, and the Hard Mobile Launcher beingdeveloped for the Small ICBM (“Midgetman”),lightsat launchers could pursue survivabilitythrough mobility. They could also employ conceal-ment, as do the Pershing 2 launcher and submarines(e.g. Trident) that launch ballistic missiles.

Operational military lightsat launchers wouldprobably be required to launch on short notice and tosustain higher launch rates than typical space launchfacilities do. On the other hand, lightsat launchvehicles would be useful with less lift capability thanmost launch vehicles have, although DARPA hassaid the Scout launch vehicle lifts too little to beuseful as an interim launch vehicle for launchingdevelopmental lightsats, and the Scout does not havethe survivability and launch rate desired for wartimeuse. The Pegasus launch vehicle may provide aninnovative means of improving launch flexibilityand survivability.

Coherent of Modular Imaging Collectors (COSMIC) described by the National Research Council in Space Technologyto Meet Future Needs (Washington, DC: National Academy Press, 1987), p. 39.

“Sub-millimeter Waves,“ in & Micro Spacecraft for Space Science Workshop-Presentations, CaliforniaInstitute of Technology Jet Propulsion Laboratory, July 6-7, 1988; pp. 136-142.


Figure 4-4-Orbiting Low-Frequency Array of Radioastronomy Satellites (Artist’s Concept)

~ “ l o o ‘ M

SOURCE: California Institute of Technology, Jet Propulsion Laboratory.

Civil lightsats and developmental militarylightsats would not require transportable launchers.

LAUNCH SYSTEM OPTIONS

Existing Launch Vehicles

Civil lightsats, or developmental military lightsats,could be launched-alone or co-manifested withother payloads-on currently operational launch

vehicles larger than the Scout. If they are smallenough, they could be launched by the Scout.

Air-Launched Vehicles

The communications satellites now being devel-oped by DARPA are very light; several can belaunched on the Pegasus air-launched vehicle (ALV).The Pegasus ALV is being developed by OrbitalSciences Corp. (OSC) and Hercules AerospaceCorp. as a $50 million privately funded joint


Figure 4-5-Pegasus Launch

The Pegasus air-launched vehicle is released from a modified B-52 aircraft operated by NASA’s Dryden Flight Research Facility (artist’sconcept).SOURCE: Orbital Sciences Corp.

venture. DARPA will pay OSC $6.3 million toprovide the launch vehicle for a government demon-stration launch. This price includes neither the costof using NASA’s B-52 as an ALV carrier (see figure4-5) nor the cost of safety support from the AirForce’s Western Test Range. Pegasus is expected tobe able to launch a 335-kg (738-lb) payload into anorbit 500 km (270 nmi) high and inclined 25 degrees,or a 244-kg (537-lb) payload into a polar orbit 500km high (see figure 4-6).

DARPA originally intended to launch the Micro-sats on the first Pegasus launch, but has decided tolaunch them on the second launch, perhaps late in1989. On its first flight, now scheduled for January1990, Pegasus will carry:

1. a 150-lb Navy communication satellite,2. a NASA scientific experiment payload, and

3. instrumentation to evaluate the performance ofthe ALV.

DARPA’s contract has four launch options re-maining, and OSC has offered to add six additionallaunch options to the contract.

OSC and Hercules also expect Pegasus can launchlightsats into geostationary orbit; they recentlysigned an agreement with Ball Aerospace to launchtwo BGS-1OO Ball geostationary satellites in late1990 or early 1991. OSC estimates each launch willcost between $6 million and $8 million, and thateach of the satellites, which will weigh about 400 lband carry a 400-watt single-channel transponder,will cost between $5 million and $8 million. OSCestimates each mission (satellite, launch, and sup-port) will cost about $20 million. The satellite design


Figure 4-6-Projected Performance of Pegasus: Payload v. Orbital Altitude and Inclination

1 Perigee Altitude, nmi

450

\

350

\

250 150

\ \ / \● ● ●

\ .

\ \

\ \ \ 1. *%%% ] \ \ 4 50 \ 3 53, 23. ,),.

nm

500

9

Circular Orbits

r t) I I I I I I I I

o 300 400 500 600 700 800 900 1000

Payload to Orbit, lb

E q u a t o r i a l o r b i t s— “ — O “ i n c l i n a t i o n

SOURCE: Orbital Sciences Corp.

is modular; Ball is developing larger versions withmore transponders or more power per transponder. 15

Standard Small Launch Vehicle

Last year, DARPA issued a Request for Proposalsto develop a transportable ground-launched Stan-dard Small Launch Vehicle (SSLV) capable oflaunching 1,000 lb of payload to a polar orbit 400nmi high. DARPA recently awarded a contract toSpace Data Corp. (a subsidiary of OSC) for develop-ment, one launch (from Vandenberg Air Force Basein the fall of 1991), and options for four more. Thefirst stage of the vehicle proposed by Space DataCorp. will use a solid rocket motor developed for thePeacekeeper ICBM by Morton Thiokol. The second,third, and fourth stages will be the first, second, andthird stages of Pegasus, without the wings. OSC is

Polar o r b i t s90” inclination

also developing a commercial version of the SSLV,called Taurus.

Other small launch vehicles have been developedor proposed by companies that performed Phase 1SSLV studies for DARPA. Space Services, Inc.(SSI) developed a launch vehicle called Conestoga,which uses clustered Castor solid rocket motors. Thefirst Conestoga was successful on a sub-orbitalflight, but the second failed shortly after launch onNovember 15, 1989. LTV Aerospace, which pro-duces the Scout, could produce an upgraded version.Lockheed Missiles and Space Co. proposed a launchvehicle that would use the first- and second-stagemotors from Poseidon C3 fleet ballistic missiles andMorton Thiokol Star 48 motors for the third stage.Lockheed estimated that the vehicle could beavailable in two years and could launch a 770-lb

Is’’ Pegasus, Ball to Launch Communication Satellites Into Geosynchronous Orbit,” Aviation Week and Space Technology, June 12,1989, p. 64, andMilitary Space, June 19, 1989, p. 8.


payload from a land-based launcher to a 250-milehigh orbit inclined 28 degrees.16

Other Options

The U.S. Naval Research Laboratory (NRL) isdeveloping a concept for a sea-launch system tolaunch a partially submerged launch vehicle from aplatform towed out to sea.17 This system, calledSEALAR (SEA Launch And Recovery) mightprovide the survivability required by lightsats.Sea-launch systems have been tested, to differentdegrees, by the U.S. Navy (Project Hydra), TruaxEngineering 18 (SEA DRAGON, SUBCALIBER),and StarStruck (now American Rocket Co.). DARPAis not known to be considering a sea-launch systemfor its Advanced Satellite Technology Program, butsuch a system might prove attractive to the Navy inthe future.

Storage, shipment, and mobile basing of smalllaunch vehicles could be made safer by using hybridrocket motors—rocket motors that use liquid oxy-gen to burn solid fuel, which can be inert (nonex-plosive). American Rocket Co. (AMROC) of Cam-arillo, CA, has developed a throttleable, restartable,70,000-lb thrust hybrid rocket motor, the H-500. Onits first launch attempt (October 5, 1989), the motorfailed and the prototype sounding rocket it was topower collapsed and burned at Vandenberg AirForce Base, CA. It is noteworthy that it did notexplode, and did very little damage to the pad.

The sounding rocket, a prototype of AMROC’Splanned Industrial Research Rocket, carried a pay-load designed by AMROC for a Strategic DefenseInitiative experiment and a prototype reentry vehicledeveloped by the Massachusetts Institute of Tech-nology (MIT) Space Systems Laboratory. The reen-try vehicle was to deploy an umbrella-like structuremade of space-suit material and decelerate to a softlanding in the Pacific Ocean, where it was to berecovered.

AMROC is developing a larger sounding rocketand an even larger Industrial Launch Vehicle, thelargest version of which is being designed to launcha 4,000-lb payload into low Earth orbit. AMROC is

not specifically designing the launch vehicle forsurvivable basing (although this is not precluded)and has not entered DARPA’s SSLV competition.Before the sounding rocket failure, AMROC ex-pected a frost launch late in 1990.

AMROC is also developing a larger hybrid motorfor a larger launch vehicle, as well as a smallerhybrid motor for various applications, possiblyincluding use on projectiles launched from electro-magnetic launchers (discussed below).

General Technology Systems (GTS) is develop-ing a small launch vehicle called LittLEO to launchlightsats. It is expected to be able to launch almosta tonne (2,200 lb) of payload into a polar orbit 300kilometers (162 nautical miles) high. First launch isplanned for 1992, probably from Andoya, Norway.GTS quotes a price of nine million pounds sterlingper launch, which is equivalent to roughly $6,400per pound to LE0.19

E’Prime Aerospace Corp. (EPAC) is developinga series of launch vehicles for launching payloadsweighing from 1,000 to 20,000 lb into LEO and upto 8,000 lb into geostationary Earth orbit (GEO).EPAC quotes a prices of $12 million per launch forthe smallest launch vehicle and $80 million for thelargest. 20 EPAC plans to launch from Cape Canav-eral Air Force Station, FL, and from Vandenberg AirForce Base, CA. A first launch is planned for 1992.

The Soviet Union is developing a launch vehiclecalled “Start” for launching lightsats. Start woulduse guidance and propulsion systems developed forthe SS-20 ballistic missile and could be launchedfrom a mobile launcher, carrying 300-lb payloads toorbits 500 km high. Space Commerce Corp., inHouston, is seeking customers for Technopribor,which is developing Start, and quoting a price ofabout $5 million to $6 million per launch. Tech-nopribor estimates a test launch could be conductedin 1991.

Lightsats can also be launched as “piggyback”payloads on launch vehicles carrying larger primarypayloads. Many U.S. and foreign launch vehicleshave done this for years. Arianespace has developed

lbt+bctieed will ~velop Small Milit~ Booster Using Poseidon C3 H~dw~e! “ Aviation Week and Space Technology, Sept. 26, 1988, p. 18.ITNRL, Nav~ Center for Space Technology, briefing for OTA staff, Feb. 17, 1989.

lgcapt, Robefl C. Truax, USN (ret.), “Commercial View on Launch Vehicles, ” Space Systems Productivity and Manufacturing Conference [V (ElSegundo, CA: Aerospace Corp., Aug. 11-12, 1987); pp. 55-69.

lgIan p~ker, “Getting There Cheaply, ” Space, vol. 5, No. 4, July-August 1989, pp. 45-48.zOIbid.


procedures to do so routinely with the Ariane 4launch vehicle, which, as noted above, is scheduledto launch six amateur-radio satellites in addition tothe SPOT 2 photomapping satellite on January 19,1990. General Dynamics is planning to offer asimilar service using its commercial Atlas launchvehicle, which could launch, in addition to a primarypayload, one 3,000-lb satellite or several smallerlightsats to LEO, or a 2,000-lb payload to geosta-tionary transfer orbit.21 This service could be offeredin late 1991 for about $6,000 per pound to LEO; theprimary payload owner may reserve the right toapprove the price offered.22

Someday, small lightsats might be launched onlaser-powered rockets (discussed below). Lightsatscould also be launched on vehicles proposed forlaunching larger payloads or crews--e.g., the Ad-vanced Launch System, the Advanced MannedLaunch System, and NASP-derived vehicles.23 Ofthese, only National Aero-Space Plane (NASP)-derived vehicles (NDVs) are intended to provide asurvivable capability for wartime launch.

ISSUESAre lightsats the most economical answer to the

problem of satellite vulnerability? Replenishingsatellites in wartime is only one of several partialsolutions; others include hardening satellites andstockpiling spare satellites in orbit during peace-time,24 as well as arms control, actively defendingsatellites, and reducing reliance on satellites forsupport of military operations.25

What military requirements could lightsats sat-isfy? An Air Force officer responsible for spacesystem planning said, “The challenge to the small

satellite community has been to get out of the moldof a solution looking for a problem; that is, whatmissions will a small satellite support. "26 Accordingto the previous Secretary of the Air Force, “Thedecision on whether a system is ‘small’ depends onsuch things as orbit, mission, requirements, andtechnology capabilities. When these factors properlyconverge, we have built Smallsats . . . . What wewant are a realistic set of requirements and conceptsfor smaller systems.”27 Several concepts havealready been proposed, the most grandiose of whichare being considered by the Strategic DefenseInitiative Organization: ‘‘BrilliantPebbles” or largerspace-based missile interceptors, “Brilliant Eyes”(space-based space-surveillance satellites that woulddemonstrate Brilliant Pebbles technology—see box4-A), decoys for Brilliant Pebbles, and “SmallDumb Boosters” (orbital transfer stages with whichBrilliant Pebbles could rendezvous and mate).

Could lightsat technology and launch vehiclesbenefit civil applications? They already have. Forexample, for two decades amateur radio operatorshave built and used lightsats for recreational, educa-tional, and public-service communications. Con-ceivably, networks of tens or hundreds of lightsatscould provide continuous global communications ornavigation services commercially. There is somecommercial interest in concepts that would requireonly a few lightsats.2829

Since Sputnik I and Explorer I, science hasbenefitted from lightsats and will continue tobenefit-more so if launch costs are reduced.Medium-sized multi-mission remote-sensing satel-lites have used some instruments, such as theScanning Multichannel Microwave Radiometer onSeasat-1, 30 that are light enough to be mated to a

zlTtJ reach g~~tionq orbit (GEO) from geostation~ transfer orbit (GTO), the payload must have an orbital wansfer stage.

zzIan pinker, op. cit., footnote 19.

23s= U.S. Cmgess, ~fice of T~hnology Aswssment, Round Trip to Orbit: Human SpaceflightAlternatives, OTA-lSC-419 (W*in$On, DC: U-s.Government Printing Office, August 1989).

24sW, for exmple, tie sp~ch hat tie Honorable Edward C. AMridge, Jr., Secretary of the Air Force, prepared for a luncheon at the Aviation club>Crystal City, VA, Sept. 15,1988, and Col. Charles Heimach, USAF, Speech to Second Annual AIAA/USU Conference on Small Satellites, Utah StateUniversity, Logan, UT September 1988.

2S’S= U.S. Cm=ss, office of T~hnology AsSssment, Antisatellite Weapons, Countermeasures, and Am Cotirol, OTA-ISC-281 (w’sshin~on~DC: U.S. Government Printing Office, September 1985); reprinted in Office of Technology Assessment, Strategic Defenses (Princeton, NJ: PrincetonUniversity Press, 1986).

Zwol. Chmlm Heimach, USAF, op. cit., footnote 24.

Z7Hm. Edwmd C. Aldridge, Jr., op. cit., footnote 24.

zgF~ and Mosier, op. cit., footnote 1.2~e ~n~r for ~ovative TWhnoloa of (he Comonwedth of Virginia h~ commissioned ~onomi~s at G~rge Mason University to ~StXS

potential markets for small satellites. This study is nearing completion.

30A.R. Hibbs and W.S. Wilson, “Satellites Map the Oceans, ” IEEE Spectrum, October 1983, pp. 46-53.


Box 4-A—Brilliant Eyes?Lawrence Livermore National Laboratory (LLNL) is developing a new class of electronic high-resolution

wide-angle TV cameras that, from an altitude of 1,000 km (610 mi), could image a land area the size of the stateof Virginia and show individual buildings. The first prototype camera, completed in 1987, has optics that are about1 ft in diameter and 16 in long, excluding electronics. With improved electro-optical components, its resolutionwould be comparable to or slightly better than that of the French SPOT satellite (about 10 m) from a comparablealtitude (832 km). At a lower altitude it could show greater detail but would have a smaller field of view. On Earth,it could be used as a telescopic TV camera to record the tracks of meteors and low-altitude satellites against the

night-time sky.SPOT Imagery of the Pentagon and LLNL has also developed a preliminary

the White HouseLightsats Could Produce Comparable Imagery

design for a miniature version of this cameracompact enough for use as a satellite naviga-tion system. The system is designed to getperiodic position updates by viewing manystars at the same time. The total mass of thesystem is expected to be less than 250 grams(about half a pound). LLNL expects that “this[wide-field-of-view] system, with its combi-nation of high resolution and high light collec-tion capability, will also find applications in

SOURCE: U.S. Geodetic Survey. robot vision and smart munitions.

1LLNL, Energy and Technology Review, July-August 1988, pp. 88-89.

lightsat bus to become a lightsat for meteorology and mapping or “mediasat” applications.31 This mightoceanography. Arrays of lightsats could someday be more economical than using a large, monolithicuse interferometric (phased-array) and aperture-satellite; predicting an arrays’ relative economysynthesis techniques to provide high-resolution would require comparing cost estimates based onradar or microwave imagery for Earth-resources detailed designs.

U.S. Congress, Office of Technology Assessment, Commercial From Space, (Springfield, National Information Service, May 1987).

Chapter 5

Microspacecraft

Microspacecraft would be satellites or deep-spaceprobes weighing no more than about 10 kilograms(22 pounds).1 Tens or hundreds could be usedtomeasure magnetism, gravity, or solar wind at widelyseparated points simultaneously. A swarm of differ-ent microspacecraft could obtain detailed radioimages of galaxies, while others could be used forcommunications, gamma-ray astronomy, or plane-tary photoreconnaissance.

They would not require development of newlaunch systems; they could be launched like buck-shot on existing small launch vehicles. However, ifthere were a demand for launching thousands peryear, it might be cheaper to launch them onlaser-powered rockets (see figure 5-1 ), if these provefeasible.

Extremely rugged microspacecraft, constructedlike the Lightweight Exe-Atmospheric Projectile(LEAP) being developed for the Strategic DefenseInitiative,2 could be launched to orbit by an electro-magnetic launcher (railgun or coilgun) or a ramcannon. An electromagnetic launcher in orbit couldlaunch them toward outer planets at muzzle veloci-ties that would allow them to reach their destinationsand return data to Earth within a few years. Thismight allow a graduate student to design a missionand then receive mission data in time to use it in aPh.D. dissertation.

SPACECRAFT CONCEPTSSeveral concepts for microspacecraft have been

proposed. One example is the Mars ObserverCamera (MOC) microspacecraft proposed by the JetPropulsion Laboratory at the California Institute ofTechnology (CalTech). It would be a generic imag-ing microspacecraft; dozens could be launched oneach of several missions-to the Moon, the planets,their moons, comets, and asteroids. A MOC micro-spacecraft would be shaped like an oversized hockeypuck, about 15 centimeters (cm) in diameter and 4cm thick (see figure 5-2). It would weigh about 800grams (g). A version could be designed to withstand

the accelerations to which electromagnetic launchwould subject them.

Placed in different orbits or trajectories, theycould trade off field of view for resolution, or viceversa. For example, one MOC microspacecraft in apolar orbit about Mars could serve as a Martianweather satellite, providing two-color images with aresolution of 5 to 10 kilometers (km )-sufficient toresolve Martian clouds. A similar MOC microspace-craft in a lower orbit could serve as a mapper,providing two-color images of a smaller field ofview with better resolution—100 meters (m). In timeit could map the entire planet. A similar MOCmicrospacecraft in an even lower orbit about theMoon could provide a two-color global map of theMoon with 10 m resolution. Existing global maps ofthe Moon currently show no features smaller thanseveral hundred meters.

LAUNCH SYSTEM OPTIONSNew, specialized launch systems need not be

developed to launch microspacecraft, because theycould be launched on existing launch vehicles—bythe dozens, if appropriate. However, some proposedunconventional launch systems might prove to bebetter or cheaper than conventional launch vehiclesfor launching microspacecraft.

One example of such a system is a laser-poweredrocket that would use a laser beam, instead ofcombustion, to heat the propellant, which could beinert (i.e., nonreactive). If feasibility is proven, a10-megawatt (MW) laser may be able to launch al-kg payload of one or more microspacecraft; agigawatt laser might launch a l-tonne (t) payloadconsisting of several microspacecraft or a largerspacecraft.

The smallness of microspacecraft has anotherpotential advantage: some microspacecraft could bebuilt to withstand high accelerations comparable tothose endured by cannon-launched guided projec-tiles such as Copperhead.3 Such “g-hardened”

IS= ROSS M. Jones, “Coffee-can-sized spacecraft,” Aerospace America, October 1988, pp. 36-38, and “Think small-in large numbers, ”Aerospace America, October 1989, pp. 14-17.

Zsee U.S. Congess, Office of TwhnoIogy Assessment, SD1: Techncdogy, Survivubiliry, and So@are, OTA-l$C-353 (Washingon, DC: U.S.Government Printing Office, May 1988), pp. 120-121.

sThe U.S. tiy’s M712 Copperhead cannon-launchd guided projectile is an artillery shell fired from a 155mm howitzer. It hm a %nsor andelectronics for detecting a spot of laser light on a target illuminated by a low-power, pulse-coded, target-designating laser aimed by a soldier or pilot.When Copperhead detects such a spot, it steers toward it using fins deployed after launch. Copperhead rounds cost about $35,000 each.

–29–


Pitch-over

Figure 5-1—Laser-Powered Rockets

\ \‘ - ’ .

——

Payload

Ascent

Fairing (nose cone)— B a s e l a t e

lPropellant(plastic,

~ L a s e r b e a m

ice, or metal)

.A–Laser(nOtiOna l)\\

\ .‘-’--..,.


microspacecraft could be launched by “directlaunch” systems4 such as railguns,s coilguns,6 and

Orbitalinsertion

A simple laser-powered rocket would have four parts: thepropellant (a block of plastic, metal, or ice), the payload,the payload fairing (nose cone), and a base plate to whichthe propellant is bonded and the payload and fairing areattached. A large laser built on a mountain would beampower to the rocket, vaporizing the propellant andthereby producing thrust perpendicular to the surface ofthe propellant.

The figure at left illustrates operation of the systemshortly after launch, when the rocket ascends vertically.After the rocket rises above the densest part of theatmosphere, the laser beam would be aimed off-center toproduce thrust asymmetrically, causing the rocket to pitch(tilt) over, as shown above left. The rocket then begins toaccelerate downrange while ascending. When the rocketreaches orbital altitude, it continues to acceleratehorizontally, as shown above, until orbital velocity isattained.

ram cannons.7 In the near term, microspacecraftcould be launched by chemical rockets, such as

(or a mission payload), no after they leave the muzzle of the launch system, if launched into solar orbit or an interplanetary trajectory. A projectile would require a small rocketmotor or some other kind of motor to enter Earth orbit.

R. and Ali E. Dabiri, “Electromagnetic Space Launch: A in Light of Current Technology and Launch Needs andFeasibility of a Near-Term Demonstration,”IEEE Transactions on vol. No. 1, January 1989, p. 393ff.

and “Basic Principles of Coaxial Launch Technology,’ IEEE Transactions on vol. No. 2, March1984, pp. 227-230.

for of IAF-87-211, presented at the 38th

Congress of the International Astronautical Federation, Brighton, England, Oct. 10-17, 1987.

Chapter 5--Microspacecraft ● 31

Figure 5-2-Mars Observer Camera microspacecraftDesign Cutaway View-Actual Size

SOURCE: Jet Propulsion Laboratory, 1988

Scout and Pegasus, that are designed to launchpayloads of a few hundred kilograms for a fractionof the cost of launching them on larger launchvehicles.

In the remainder of this section we focus onunconventional launch technologies. A surprisinglylarge number of them have been proposed; onerecent review8 lists 60 propulsion technologies—inaddition to conventional chemical rocket technology—that are potentially applicable to space transporta-tion. About half are applicable to Earth-to-orbittransportation; most are applicable to in-spacetransportation (e.g., orbital transfer or escape),which demands less thrust and power than does

Earth-to-orbit transportation. For brevity, we discussonly two unconventional launch technologies here:railguns and two-pulse laser-supported-detonation(LSD) thrusters, which are the simplest of severalproposed laser-powered rockets. We discuss onlytheir application to Earth-to-orbit transportationhere, although both are also applicable to in-spacetransportation.9 In fact, orbital transferor reboost oflow-altitude satellites could be done with muchsmaller lasers than would be required for launchingprojectiles from Earth to orbit.10

Direct launch systems would subject payloads tohigh accelerations. For a specified muzzle velocity,the barrel length of any type of direct-launch systemmust grow as the reciprocal of acceleration. Toachieve a muzzle velocity of 8 kilometers per secondwith an acceleration of 1,000 gs,ll a direct-launchsystem must have a barrel more than 3 km long. Itwould be impractical for a launcher to be muchlonger, or to subject a payload to correspondinglylower acceleration.

To launch a projectile vertically at an accelerationof 1,000 gs, the projectile must be subjected to aforce 1,001 times its weight.12 Hence exotic design,fabrication, and testing processes are required—especially for electronic and optical components—and there are constraints on the shape of theprojectile, and, in practice, limits on its size.Proposed projectiles have weights ranging from afew kilograms to a tonne. They could carry payloadssuch as fuel, food, water, structural components forspace assembly, and specially designed electronicand optical systems such as those used in the Army’sCopperhead cannon-launched guided projectile andSADARM cannon-launched sensor-fuzed weapon,and the Lightweight Exe-Atmospheric Projectilebeing developed for the SDI.

Eder, “Technological Progress and Space Development,” draft, Apr. 30, 1989. e.g., Ross M. ‘‘Electromagnetically Launched Micro Spacecraft for Space Science Missions,’AIAApaper 88-0068, Jan. 11, 1988, and

Arthur Kantrowitz, ‘‘Laser Propulsion to Earth Orbit: Has Its Time Come?’ (cd.), Proceedings Workshop on Laser Propulsion, CA: Lawrence National Laboratory, July 7-18, 1986), vol. 2, pp. 1-12.

op. cit., foomote 9, and Chapman,‘‘Strategic Defense Applications of Ground-Based Laser Propulsion, ’ (cd.), op. cit.,footnote 9.

is ‘g”) of acceleration. acceleration of g, an object’s increases by 9.8 m (32 per second per second.

force, it accelerate downward; its velocity would increase by9.8 m per second each second. If sitting on the ground, it would be subjected to a force to its weight; this keeps it from falling into the ground-itaccelerates at zero If subjected to a force twice its weight, it would accelerate upward at 9.8 m per second per second, If subjected to a force thriceits weight, it would accelerate upward at 19.6 m per second per second, and on.


In a previous report, OTA described two proposeddirect launch systems: ram cannons and coilguns.l3

Coilguns could have important advantages overrailguns, which are simpler and more familiarelectromagnetic launchers (EMLs). For example,coilguns can be designed so that the projectile doesnot contact the barrel, avoiding barrel erosion, andthey can be scaled to launch large masses efficientlyat high velocities. Until recently, railguns wereexpected to be very inefficient at launching multi-kilogram projectiles at the muzzle velocities (morethan eight kilometers per second) required to reachorbit with minimal assistance from rockets. Lowefficiency would cause barrel heating and melting,as well as a high electric bill. However, in recenttests, railguns demonstrated unexpectedly high effi-ciencies in accelerating small projectiles to muzzlevelocities of 3 to 4 km per second,14 raising hopesthat they might be able to accelerate half-tonneprojectiles to more than 8 km per second.

Another cause for increased interest in railgunsfor direct launch is the realization that ordinaryautomobile batteries could be used for energystorage and would cost much less than alternativespreviously considered. Automobile batteries couldalso power coilguns.

The Air Force recently decided to demonstratesuborbital launching of a microspacecraft using arailgun at Eglin Air Force Base, Florida, afteraugmenting its battery power system and adding abarrel with a 30-cm bore, similar to one usedrecently at Maxwell Laboratories. According to oneestimate, the upgraded gun could launch 5 to 10 kgat 4 km per second three years from program start foronly about $10 million. A l-kg projectile launchedat 3 km per second with a l-kg sabot is expected toreach an altitude of 200 km if the projectile’s nosetipis allowed to ablate, or 400 km if the nosetip iscooled by transpiration. *S

Proponents predict that a prototype operationalEML capable of launching 500-kg projectiles eachcarrying about 250 kg of payload could be developedin about six more years for an additional $900million to $6 billion, including $50 million to $5billion for development of vehicles and trackingtechnology. l6 17

If produced and launched at a rate of 10,000 peryear, projectiles (less payloads) would cost as littleas $1,000 per kg according to one estimate, but over60 times this, according to another estimate. AnEML projectile would require (besides its missionpayload) guidance, navigation, and control systems,as well as a rocket kick motor to inject it into Earthorbit. 18 Just as allowing a payload of specifiedfunction to be larger allows it to be cheaper,miniaturizing it makes it more costly, and g-gardening it would make it still more costly. On theother hand, if a mission required many projectiles,high-rate production and learning effects couldreduce unit costs. Other launch costs might be as lowas $50 per kg at this rate.

If batteries are used, the limit would be about10,000 launches (2,500 t) per year.19 Because of thebrief launch windows for rendezvous, very little ofthis tonnage could go to a space station,20 but mostor all could be used for other applications, e.g.,distributed low-altitude networks of tiny satellitesfor communications, space surveillance, ballistic-missile defense, or space defense.

If the payload were reduced and the projectiles’chemical rockets enlarged, the projectile would havemore cross-range capability; the launch windows forrendezvous with a space station would be longer,and more payload could be delivered to a spacestation per year. However, the launch cost per poundwould increase. Chemical rockets could also be usedto reduce the muzzle velocity required, so that a

13u.s. Congess, Office of TechrIoIogy Aswssment, Launch Options for the Future: A Buyer’s Guide, OTA-ISC-383 (Washington, DC: U.S.Government Printing Office, July 1988), p. 50.

~q~les R. palmer, “Electromagnetic Space Launch, ’ SA88091 (Mchan, VA: Science Applications International Corp., May 6, 1988).IsIbid.161bid,

ITMiles R. p~mer, perWn~ communication, Oct. 12, 1989.ISA space pro~ bo~d for SOlaI orbit or a fly-by of another planet would not rwuire ~is.

19A ~eater la~ch rate could be achieved at greater cost if batteries are not u~d.z(tp~mer ~d Dabiri, op. cit., fOOtnOte 5.


smaller 21 or more conventiona122 direct-launch sys-tem could be used.

Laser propulsion would have important advan-tages over direct launch. Acceleration would bemuch lower—about 6 or 7 gs on typical trajectoriesto low orbit—so payloads would not have to bedesigned to withstand gun-like stresses. Moreover,no expensive device to store and quickly dischargegigajoules of energy would be required, as it wouldbe for a railgun or coilgun, because power would bebeamed to a rocket continuously during ascent.Perhaps the most important advantage of laserpropulsion is that a simple laser-powered rocket,unlike an EML projectile, would not require guid-ance, navigation, and control systems, or a separatekick motor for injection into orbit.

But laser propulsion would require a powerful,expensive laser and a large, expensive adaptivemirror. For efficient utilization and low average costper launch, both must operate reliably withoutmaintenance for longer periods than do existinglasers. And laser launch operations would be haltedby overcast that would not impede direct launch.Moreover, laser propulsion technology is less devel-oped than EML technology and is predicated onunproven theories of thermal blooming suppressionand thruster plasmadynamics. Validation of thesetheories may require construction of a full-scalelaunch system.

An SDIO official has estimated that a 20-MWcarbon-dioxide laser with a 10m-diameter beamdirector telescope could launch rockets carrying 20kg of payload for an incremental cost of about $120per lb, assuming the laser efficiency is 15 percent,the rocket efficiency is 40 percent, electricity can begenerated for four cents per kilowatt-hour, and thestructure and propellant each cost about as much asthe electricity. According to the SDIO, the 2m-diameter rocket structure would weigh only a few kgand would require only about 120 to 150 kg of inertpropellant such as ice or polyformaldehyde plastic;

launching would require 30 to 40 megawatt-hours(MW-h) of electric power.23

If the system could operate continuously withoutdowntime for maintenance or overcast, the launchrate could be almost 60,000 per year. In practice,occasional overcast would make full utilizationunachievable, and approaching it would probablyrequire at least two lasers and mirrors, so that onepair could operate while the other is being serviced.Even this would not assure operation most of thetime, unless the duty cycle of the lasers (i.e., thefraction of time they are lasing) is much greater thanthe duty cycles demonstrated by industrial and otherhigh-power lasers (see box 5-A). However, if alaunch rate of 100 per day could be maintained, over1,600,000 pounds of payload could be launched intoorbit each year. This would be almost twice theestimated combined capability of current U.S. spacelaunch systems,

24 almost three times the annualtonnage launched in 1984 and 1985, and about fourtimes the annual tonnage launched from 1980 to1985.25

The SDIO postulates that, in practice, 100 pay-loads could be launched per day, on the average, andthe average cost could be as low as about $200 perpound, if capital cost (table 5-1) were depreciatedover 5 years and if annual operating cost (excludingrocket cost) were comparable to the annualizedcapital cost of $90 million. The SDIO estimates thatlaunching only one or two payloads a day (500 peryear) would be sufficient to reduce average cost toabout $4,500 per pound and make laser-poweredrockets competitive with conventional rockets forsmall payloads. Some users might be willing to paya premium for the speed with which a laser-poweredrocket could be prepared to launch a payload.

The SDIO estimates that a first launch to orbitcould be attempted about 5 or 6 years after programstart and expects to demonstrate a rocket efficiencyof 20 percent or more in experiments now beingplanned. However, the highest efficiency demon-

zlp~mer and Dabiri, op. cit., footnote 5, and Henry Kolm and Peter Mongeau, “An Alternative Launching Medium,” IEEE Spectrum, vol. 19, No.4, April 1982, pp. 30-36.

Z?L~ge ~amOn~ have &aen ~@ tO ]a~~h ~ubo~bltal ~rOjectlle~ and sounding rockets to altitudes of 4oo,0~ feet; sw C.H. M~hy a n d G.V. Bu1l,“AReviewof Project HARP,” Annah of theiVew York Academy of Science, vol. 140, No. 1, 1966, pp. 337-357, and R.G.V. Bull and Charles Murphy,Paris CannonetiThe Paris Guns and Project HARP (Springer-Verlag, November 1988).

23 Jordin T. K=e, ~~~ls~ La~r ~opulsion for ~w Cost, High Volume La~ch to mbit,” preprint UCRL-101 139, Lawrence Livermore National

Laboratory, Livermore, CA, June 2, 1989. A different launch simulation program (Kantrowitz, op. cit., footnote 9.) predicts each launch would requireabout nine minutes (hence about 20 MW-h of electric power) and about 200 kg of propellant.

XOTA, hunch optw~ for tk Future: A Buyer’s Guide, op. cit., footnote 13, p. 20.

~bid., p. 5.

34 . Affordable Spacecraft: Design and Launch Alternatives

Box 5-A—Lasers for Rocket Propulsion: The State of the ArtA laser-powered rocket would use a laser beam, instead of combustion, to heat the propellant, which could be inert

(i.e., noncombustible). The beam could come from a ground-based laser; the rocket could be extremely simple and weighonly 10 times its payload. For comparison, the Scout launch vehicle weighs 1,300 times its payload.

Studies of Earth-to-orbit laser propulsion postulate the use of infrared lasers, which could be carbon-dioxide ordeuterium-fluoride electric-discharge lasers or free-electron lasers. The most mature of these is the carbon-dioxideelectric-discharge laser, but free-electron lasers are more efficient. If a carbon-dioxide laser or a free-electron laseroperating at the same wavelength (0.01 mm) were used, the economical laser power would be about 1 megawatt (MW:1 million watts) per kilogram (kg) of payload, if the laser-powered rockets can achieve the 40-percent energy-conversionefficiency once predicted by laser-propulsion proponents. To date, only 10-percent efficiency has been achieved.Laser-propulsion experts now predict that at least 20-percent efficiency can been achieved. At this efficiency, a laser powerof about 25 MW might be required to launch a 20-kg payload. If, pessimistically, no more than 10-percent efficiency isachieved, about 50 MW might be required to launch a 20-kg payload.

It appears to be feasible to build such a laser; the U.S. has built a gigawatt (billion-watt) free-electron maser andelectric-discharge lasers of much greater peak power, but none that could produce even 10 MW average power for tenminutes, the boost duration required to reach low orbit. Almost a decade ago, the Antares carbon-dioxide electric-dischargelaser at the Los Alamos National Laboratory produced brief pulses with a peak power of 40 terawatts (40 trillion watts).But a very different design, similar to that of industrial lasers used for welding, would be required for prolonged operationat high average power. Free-electron lasers have, to date, produced less peak power. A free-electron laser developed byLos Alamos and Boeing has produced pulses of ten megawatts peak power, but only six kilowatts average power, at awavelength of 0.01 mm. In early experiments, the partially completed Paladin free-electron laser at the LawrenceLivermore National Laboratory amplified five-megawatt pulses from a carbon-dioxide laser 500-fold, presumablyproducing pulses of about 2.5 gigawatts peak power. The carbon-dioxide laser power is being increased to a gigawatt, andthe free-electron laser has now been extended. If its electron accelerator operates at the average power for which it wasdesigned (at least 25 megawatts), and if 40 percent of the electron-beam power is converted to laser beam power(comparable to the efficiency demonstrated by a similar free-electron laser at a wavelength of 8.8 mm), the Paladinfree-electron laser would produce a laser beam of at least 10 megawatts average power.

Neither carbon-dioxide lasers nor free-electron lasers have demonstrated the duty cycle (the fraction of time a deviceoperates) that would be required for an operational launch system. The duty cycle of a free-electron laser designed for ahigh duty cycle would be limited primarily by the lifetime of the cathode used by the electron accelerator. Looselyspeaking, the cathode is like the filament of a light bulb, and more closely resembles the cathode of a cathode-ray tube suchas a TV picture tube. Several cathodes designed for long life are being tested. Alternatively, an electron storage ring (anarrangement of magnets) could be used to recirculate the electron beam, as was done in the first free-electron laser andothers.

Focusing a multimegawatt laser beam on a small rocket hundreds of kilometers away is another serious technologicalproblem; in particular, control of beam-degrading nonlinear optical effects, such as thermal blooming, has not yet beendemonstrated at any average power and beam diameter of interest. Some research sponsored by the Strategic DefenseInitiative Organization (SDIO) is aimed at demonstrating high-power beam control for ballistic missile defense; the beamcontrol required for propulsion would be more difficult in some respects.

Nevertheless, participants at a 1986 workshop on laser propulsion sponsored by SDIO and the Defense AdvancedResearch Projects Agency expressed optimism that a free-electron laser and beam director then planned for other purposes“should be capable of launching test payloads to low [Earth] orbit in the early 1990s. ” SDIO subsequently establisheda laser propulsion program and considered using a free-electron laser and a beam director to be developed for the SDIFree-Electron Laser Technology Integration Experiment (FEL TIE) to experiment with laser propulsion, even though theFEL TIE laser would be designed to operate at a wavelength shorter than optimal for laser propulsion.

Subsequent budget cutbacks postponed by at least two or three years the date by which the FEL TIE laser and beamdirector could be operating. More recently, SDIO decided the FEL TIE laser should use a radio-frequency linear accelerator(RF linac) similar to the one developed for the Los Alamos-Boeing free-electron laser instead of an induction linac similarto the one used in the Paladin free-electron laser. The Los Alamos-Boeing RF linac produced an electron beam of higherquality than that produced by the induction linac used by the Paladin laser; however, use of an RF linac may cause theFEL-TIE laser to produce laser pulses with a waveform that is far from optimal for laser propulsion. This, together withthe nonoptimality of the FEL TIE wavelength, may lead SDIO to abandon hope of using the FEL TIE laser for laserpropulsion experiments and force SDIO, perhaps teamed with other sponsors, to develop a laser and beam directorspecifically for laser propulsion experiments.


Table 5-l-Estimated Cost of a 20-Megawatt Laserfor Powering Rockets

Development. . . . .. .. .. .. .. .. .. ... ... ......$ 75 million

Laser . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .“$185million a

Telescope . . . . . . . . . . . . . . . . . . . . . . . . .. ... ..$lO0 millionAdaptive optics .. .. .. .. .. .. .. ... ... ... ....$ 15 millionTracking .. .. .. .. .. .. .. .. .. .. ... ... ......$ 50 millionPower plant.... .. .. .. .. .. .. .. ... ... ......$ 50 millionStructure .. .. .. .. .. .. .. .. .. .. ... ... ......$ 50 million

Total capital cost . . . . . . . . . . . . . . . . . .. ... ... $450 millionTotal nonrecurring cost . . . . . . . . . . . . . . . . . $525 million

There is a class of science and explorationmissions that can be enabled by microspacecraft(i.e., infeasible with larger spacecraft). This class ofmissions requires many simultaneous measurementsdisplaced imposition .. .. Examples . . . include: 1)a global network of surface or atmospheric sensorson planets such as Mars. . ., 2) measuring the spatialand temporal structure of magnetospheres about theEarth, Sun, or other regions of space, and 3) usingmicrospacecraftas distributed arrays for either radioor optical signals.

estimated as $25 miliion +$8 per watt.

SOURCE: Strategic Defense Initiative Organization.

strated to date is about 10 percent. If only 10 or 20percent efficiency could be attained, an 80- or40-megawatt laser would be needed, and averagecost per pound would be greater than indicatedabove. If the cost of the power plant increases inproportion to its power, average cost would be about$490 or $275 per pound for 10 or 20 percent rocketefficiency, respectively, at a launch rate of 100 perday.26

Because of the brief launch windows for rendez-vous only 2 payloads per day could be launcheddirectly to a rendezvous with the space station.Payloads launched at other times would take longerand require more fuel to rendezvous.27 The SDIOconsiders 8 payloads per day to be a conservativeestimate of the number of payloads that could belaunched to rendezvous with the space station eachday. With additional investment, the laser androckets could be given more crossrange capability.This could be done by making the beam director androckets larger or by adding a conventional chemicalrocket to the laser-powered rocket.

ISSUESWhat could microspacecraft do that conventional

spacecraft couldn’t? The consensus of the NASA/SDIO microspacecraft for Space Science WorkshopPanel 28 was that:

They would have another advantage: they couldbe launched from Earth orbit toward outer planetsby space-based electromagnetic launchers (railgunsor coilguns) at muzzle velocities that would allowthem to reach their destinations and return data toEarth years earlier than could spacecraft launched byconventional rockets. This would accelerate thecycle of acquiring knowledge.

What is the market for such services? How muchis now spent on conventional spacecraft for spacescience which microspacecraft could do? The 1988NASA budget was about $9 billion, of which about$1.6 billion was for “space science and applica-tions." 29 30

Much of this is for NASA’s ‘‘great.observatories,’ such as the Hubble Space Telescopeand the Advanced X-ray Astrophysics Facility, andfor planetary probes such as Galileo. The consensusof the NASA/SDIO microspacecraft for SpaceScience Workshop Panel was that:

microspacecraft cannot achieve the science objec-tives of the great observatory missions such as theHubble Space Telescope or the Advanced X-rayAstrophysics Facility. Also, intensive, multi-facetedscience investigations such as those of Galileo atJupiter cannot be supported by the microspacecraftconcept. . . . many space science missions will haveto continue to use established technology. Micro-spacecraft, if they are to be used in deep-spacemissions, must establish a new inheritance chain, forexample by being used in near-Earth scientific ornon-scientific missions.

z~alc~ated by OTA using the launch simulation program of Kantrowitz, oP. cit., footnote 9. The launch simulation program used by Kare,‘‘Trajectory Simulation for Laser Launching,’ Kare, op. cit. (footnote 10), pp. 61-77, predicts a 50 to 100 percent longer ascent than does Kantrowitz’sprogram (for the case of 40 percent thruster efficiency) and hence 50 to 100 percent higher electric power usage and incremental cost.

zTSee p.K. Chapman, op. cit., footnote 10.Z8NAS~0AST & SD1OflST, M1cro~Pacecra~for Space Science work.s~~eport O$rhe wor~~p Panel, Cdifomia Institute of Technology Jet

Propulsion Laboratory, Oct. 6, 1988.29UCS0 Congess, Conwessiond Budget Office, The NASA progr~ in rhe 1990’s aria’ Beyo~ (W~in@on, DC: Congressional Budget OffiCe, May

1988); figure 1; see also figure 4 and box 3.So’r’he ~p~ment of Defense space program also includes some focused space science projecw.


An EML, ram cannon, or laser (to power rockets)may permit microspacecraft to be launched fromEarth to orbit at low average cost, but only if utilizedefficiently. Maximum efficiency would require launch-ing on the order of 10,000 microspacecraft per year.How much would these microspacecraft cost? Couldspace science budgets pay for so many microspace-craft? If not, what other types of microspacecraftmight be launched by such a system to maintain anefficient launch rate?

Possibilities include:31

●

●

●

●

low-altitude comsattary);

networks (civil or mili-

“Brilliant Eyes, “ “Brilliant Pebbles,’ or “SmallDumb Boosters” for a strategic defense sys-tem; 32

logistics for a space station or other spaceoperations. Payloads could include structuralcomponents, fuel, armor, etc.; andintercontinental artillery.

The utility of these applications has not beenestablished. All require further analysis before theycan be used to justify developing a direct-launchsystem or a laser and laser-powered rockets. Someproposed logistics schemes appear more promisingthan others. For example, it is probably feasible tolaunch Small Dumb Boosters (orbital transfer stages)with which Brilliant Pebbles could rendezvous andmate. Some have proposed launching projectilesloaded with water, liquid oxygen, and liquid hydro-

gen toward the Space Station Freedom, but the costsof collecting and decanting them have not yet beenestimated.

The risk of satellite collisions would increasegreatly if tens of thousands of microspacecraft wereplaced in orbit, unless a means of collision warningand avoidance is developed. Existing space surveil-lance systems may be inadequate for tracking tens ofthousands of microspacecraft, although BrilliantEyes or Brilliant Pebbles could help with this.Ground-based lasers could be used to change theorbits of satellites equipped with slabs of inertpropellant, whether launched by laser or not.33

However, this may not be adequate for collisionavoidance, because such satellites may pass overpropulsion lasers only infrequently, so advancedwarning of a collision hazard would be required, butmight be costly and subject to false alarms.

Brilliant Pebbles would not require advancedwarning of a collision; they could be programmed toavert collisions by dodging approaching spacecraft.They could also be commanded to ram a nonmaneu-verable satellite (e.g., a failed Brilliant Pebble) thatposed a threat to more valuable U.S. and foreignsatellites. But a successful intercept might generatedebris and increase the long-term risk to spacecraft.Collision avoidance schemes based on other tech-nologies developed for antisatellite or ballisticmissile defense applications have been proposed;some would not generate debris.34

sl~les R. p~mer, op. cit. footnote 14.

gz’rhese were describ~ above in tie ~tion on Iightsats. Brilliant Pebbles would weigh tens of kilograms-more than the lightest microspacwrtit,and more than some lightsats, but still light enough to be launched by laser-powered rocket.

33A laser could be used to maneuver satellites much heavier than those it could launch into orbit.

340TA hm jmt bp an assessment of technologies for controhtg space debris and protecting satellites from it. The assessment Wm r~uest~ bythe Senate Committee on Commerce, Science, and Transportation, its Subcommittee on Science, Technology, and Space, and the House Committee onScience, Space, and Technology.

General Information

Contacts Within OTAOTA offices are located at 600 Pennsylvania Ave., S. E., Washington, DC.

Personnel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..................224-8713Publication Requests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..........224-89%Office of the Director . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .........224-3695Congressional and Public Affairs Office . ........................224-9241Energy, Materials, and International Security Division . ............228-6750Health and Life Sciences Division . . . . . . . . . . .....................228-6500Science, Information, and Natural Resources Division . . . . . . . . . . ...228-6750

Reports and lnformationTo obtain information on availability of published reports, studies, and

summaries, call the OTA Publication Request Line (202) 224-8996.

For information on the operation of OTA or the nature and status of ongoingassessments, write or call:

Congressional and Public Affairs OfficeOffice of Technology AssessmentU.S. Congress Washington, DC 2051O-8O25(202)224-9241

Other OTA PublicationsList of Publications.--Catalogs by subject area all of OTA’s published

reports with instructions on how to order them.

Assessment Activities.—Contains brief descriptions of recent publicationsand assessments under way, with estimated dates of completion.

Press Releases.—Announces publication of reports, staff appointments, andother newsworthy activities.

OTA Annual Report.--Details of OTA’s activities and summarizes reportspublished during the preceding year.

OTA Brochure.—"What OTA Is, What OTA Does, How OTA Works.”

Related OTA Reports

Civilian SpaceAdvanced Space Transportation Technologies Assessment

Round-Trip to Orbit: Alternatives for Human Spaceflight. OTA-ISC-419, August 1989, GPO stock #052-003-01 155-7;$5.50.Launch Options for the Future: A Buyer’s Guide. OTA-ISC-383, July 1988. NTIS order #PB 89-1 142681AS.Reducing Launch Operations Costs: New Technologies and Practices. OTA-TM-ISC-28, September 1988. GPO stocknumber #052-003-01 118-2; $4.50.Big Dumb Boosters: A Low-Cost Space Transportation Option? OTA Background Paper, February 1989. (availablefrom OTA’s International Security and Commerce Program)

Commercial Newsgathering From Space. OTA-TM-ISC-40, May 1987. NTIS order #PB 87-235 396/XAB.Space Stations and the Law: Selected Legal Issues. OTA-BP-ISC-41, September 1986. NTIS order #PB 87-118220/AS.International Cooperation and Competition in Civilian Space Activities. OTA-ISC-239, .July 1985. NTIS order #PB87-136 842/AS.

U.S.-Soviet Cooperation in Space. OTA-TM-STI-27, July 1985. NTIS order #PB 86-114 758/AS.Civilian Space Stations and the U.S. Future in Space. OTA-STI-241, November 1984. GPO stock #052-O03-O0969-2;$7.50.Remote Sensing and the Private Sector: Issues for Discussion. OTA-TM-ISC-20, March 1984. NTIS order #PB 84-180777.Salyut: Soviet Steps Toward Permanent Human Presence in Space. OTA-TM-STI-14, December 1983. NTIS order #PB84-181 437/AS.UNISPACE ’82: A Context for International Cooperation and Competition. OTA-TM-ISC-26, March 1983. NTISorder #PB 83-201848.Space Science Research in the United States. OTA-TM-STI-19, September 1982. NTIS order #PB 83-166512.Civilian Space Policy and Applications. OTA-STI-177, June 1982. NTIS order #PB 82-234444.Radiofrequency Use and Management: Impacts From the World Administrative Radio Conference of 1979.OTA-CIT-163, January 1982. NTIS order #PB 82-177536.Solar Power Satellite Systems and Issues. OTA-E-144, August 1981. NTIS order #PB 82-108846.

Military SpaceSDI: Technology, Survivability, and Software. OTA-ISC-353, May 1988. GPO stock #052-O03-O1084-4; $12.00.Anti-Satellite Weapons, Countermeasures, and Arms Control. OTA-ISC-281, September 1985. NTIS order #PB86-182 953/AS.Ballistic Missile Defense Technologies. OTA-ISC-254, September, 1985. NTIS order #PB 86-182 961/AS.Arms Control in Space. OTA-BP-ISC-28, May 1984. NTIS order #PB 84-198 209/AS.Directed Energy Missile Defense in Space. OTA-BP-ISC-26, April 1984. NTIS order #PB 84-21011 l/AS.

NOTE: Reports are available through the U.S. Government Printing Office, Superintendent of Documents, Washington, DC 20401(202) 783-3238; and the National Technical Information Service, 5285 Port Royal Road, Springfield, VA 22161, (703)487-4650.

Affordable Spacecraft: Design and Launch Alternatives · IBM Federal Systems Division George B....

Documents

Transcript of Affordable Spacecraft: Design and Launch Alternatives · IBM Federal Systems Division George B....