ANALYSIS OF HALL-EFFECT THRUSTERS AND ION ENGINES …

1

ANALYSIS OF HALL-EFFECT THRUSTERS AND ION ENGINES FORORBIT TRANSFER MISSIONS

Frank S. Gulczinski III*, Ronald A. Spores**

Propulsion DirectorateOL-AC Phillips LaboratoryEdwards AFB, CA 93524

ABSTRACT

Analytical methods were combined with actual thruster data to create a model used to predict theperformance of systems based on two types of electric propulsion thrusters, Hall-effect thrusters and ionengines, for several orbit transfer missions. Two missions were trip time constrained: a LEO-GEOtransfer and a LEO constellation transfer. Hall thrusters were able to deliver greater payload due to theirhigher overall specific power. For the power limited orbit topping mission, the choice of thruster isdependent on the user’s need. Ion engines can deliver the greatest payload due to their higher specificimpulse, but they do so at the cost of higher trip time. Study of reusable electric orbit transfer vehiclesystems indicates that they can offer payload mass gains over chemical systems, but that these gains areless than those offered by other electric propulsion transfer scenarios due to the necessity of carryingpropellant for return trips. Additionally, solar array degradation leads to increased trip time forsubsequent reusable transfers.

* Research Aerospace Engineer, Member AIAA** Group Leader, USAF Electric Propulsion Lab, Member AIAA

This paper is declared a work of the US Government and is notsubject to copyright protection in the United States.

INTRODUCTION:

The US Air Force has recently completedseveral studies to investigate the potentialadvantages of advanced space propulsion forseveral orbit transfer scenarios. The first studyinvestigated advanced propulsion concepts forexpendable orbit transfer vehicles1 andconcluded that the potential launch vehicledownsizing that resulted from the use of highspecific impulse thrusters provided significantcost savings over base line chemical launchvehicle/upper stage systems. The second studylooked at reusable advanced upper stages2 andpreliminary indications are that while thereremains the potential for launch vehicledownsizing, it is significantly reduced comparedto expendable systems. This difference waslargely due to the added propellant required toperform the round trip mission from low-earthorbit to geostationary orbit. Both studies pointedout advantages for advanced electric propulsionsystems based on xenon propellant. Theobjective of this paper is to analyze the trade-offs between Hall-effect thrusters and ion

engines as a high power propulsion system fororbit transfer missions.

Both the Hall-effect thruster and the gridded ionengine are classified as electrostatic thrustersand operate on heavy noble gases, primarilyxenon. These electric propulsion devices arecapable of specific impulses ranging fromapproximately 1500 to 4000 seconds, comparedto chemical systems which typically operate inthe range of 300 to 400 seconds.

Electric propulsion is a type of rocket propulsionfor space vehicles and satellites which utilizeselectric and/or magnetic processes to acceleratea propellant at a much higher specific impulsethan attainable using classical chemicalpropulsion. The concomitant reduction inrequired propellant mass results in increasedpayload mass capability.

The method of analysis used in this study isbased on the model developed by Messerole.3 Ithas been modified to reflect the most currentinformation on thruster development levels and

2

to allow for greater flexibility in systemcomponent variation.

There are four missions examined in this study.They are a low earth orbit to geosynchronousearth orbit (LEO-GEO) transfer, a LEO tointermediate orbit transfer for constellations ofsatellites, a geostationary orbit insertion withpartial chemical propulsion, and a reusable orbittransfer vehicle concept. These missions arerepresentative of the range of orbit transferscenarios that the Air Force presently envisionsfor an electric propulsion upper stage, and areamong those likely to be attempted over the next10 to 20 years.

ORBIT TRANSFER HARDWARE:

The key components of an electric propulsionorbit transfer vehicle are the thrusters and thepower generation sub-system. In this study, twotypes of thrusters are examined: the Hallthruster and the ion engine. Power is generatedby a concentrator solar array.

Hall-effect thrusters are a type of electrostaticthruster in which ions are generated by electronbombardment. The ions are then accelerated byan electron cloud which is held in place by amagnetic field perpendicular to the direction ofacceleration. The electron cloud is generated byan applied electric field.

Figure 1: Hall Thruster (SPT-100)

Initial work on Hall-effect thrusters began in the1960’s in the United States and the formerSoviet Union. Due to difficulties achieving thesame levels of efficiency reached by the gridded

ion engine, work ceased in the United Statesaround 1970.4 In the Soviet Union, researchinto the ion acceleration mechanism led toimprovements in efficiency and further researchand development. Two basic types of Hallthrusters were developed: the Stationary PlasmaThruster (see Figure 1) developed under theleadership of A.I. Morozov at the KurchatovInstitute and the Anode Layer Thrusterdeveloped under the leadership of A.V.Zharinov at TSNIIMASH.5 The primarydifferences between the two types are that theacceleration region of the SPT is within thethruster itself while for the ALT it is in front ofthe thruster and the lack of an accelerationchamber insulator in the ALT. This study doesnot distinguish between the various types of Hallthruster concepts with regards to performance.

Over sixty-four SPT-50 and SPT-70 units haveflown aboard Russian spacecraft, beginning withthe Meteor satellite in 19726 (the numericaldesignation in the name of a SPT is the outerdiameter of the discharge chamber inmillimeters). The first SPT-100s flew in 1994on the GALS spacecraft. Larger thrusters, theSPT-140, SPT-200, and SPT-290 haveundergone various levels of laboratorydevelopment. With the end of the Cold War,this technology became available for evaluationand use in the West. Work in the United Statesto further quantify SPT performance and flightqualify SPTs for western spacecraft has beendone primarily at the NASA Lewis ResearchCenter7 and the Jet Propulsion Laboratory.8

Space Systems Loral has developed powerprocessing units for the SPT-100 and is workingto develop higher power PPUs.

3

1500 2500 3500 4500 5500 6500

Specific Impulse (s)

0

5000

10000

15000

20000

25000

Pow

er (

W)

Thruster:RIT: 35 cmNSTAR: 30 cmDerated: 30 cmHughes: 25 cmRIT: 15 cmNAL: 14 cmHughes: 13 cm

MELCO: 12 cmIES, IPS: 12 cmRIT: 10 cmUK: 10 cmIon Engine: ModelHall Thruster: Model

Figure 2: Thruster Power versusSpecific Impulse

By examining a range of thruster sizes andoperating conditions,9,10,11 we are able to makemodeling predictions. One of the mostimportant parameters for this study is the curveof thruster power in terms of the specificimpulse (Isp). As specific impulse increases, thedischarge voltage will increase as well. TheHall thruster power is modeled using apolynomial fit. This relation is shown in Figure2.

The thruster efficiency is modeled using therelationship:

( )η =

+

ab

g Io sp

12

(1)

which is based on the ion thruster efficiencyequation developed by Brophy.12 This efficiencymodel is shown in Figure 3.

1500 2000 2500 3000 3500 4000


0.4

0.45

0.5

0.55

0.6

0.65

0.7

0.75

0.8

Eff

icie

ncy

(-)

Hall Thruster EfficiencyIon Engine Efficiency

Figure 3: Thruster Efficiency versusSpecific Impulse

In the gridded ion engine (see Figure 4),propellant is injected into an ionization chamberand ionized by electron bombardment. Thepropellant is then electrostatically acceleratedthrough a series of biased grids. Traditionally,these grids have been molybdenum, thoughrecent work has been done to develop carbon-carbon composite grids.

In the United States, ion engines were developedat NASA’s Lewis Research Center in the late1950’s under the guidance of Dr. HaroldKaufman.13 The original models used primarilymercury or cesium for propellants. Thrusterdevelopment has continued at various levels,using thrusters with diameters ranging from 2.5to 150 centimeters and power levels rangingfrom 50 W to 200 kW. Flight experiments haveincluded SNAPSHOT, a US Air Force satellitethat flew a cesium ion engine in 1965; SERT-2,a NASA satellite that flew a mercury ion enginein 1970; and ETS-3, a Japanese satellite thatalso flew a mercury ion engine in 1982.14 In the1980’s, emphasis shifted to xenon and othernoble gases because of concern over spacecraftcontamination and environmental issues duringground testing. As part of its New Millenniumprogram, NASA has been developing NASA’sSolar Electric Propulsion TechnologyApplication Readiness (NSTAR) engine for useon the first New Millennium mission, whichincludes an asteroid and comet fly-by. Othernations including the United Kingdom,Germany, and Japan are also currentlydeveloping ion engine technologies.15 HughesSpace and Communication Co. has undertakenan extensive development effort and hasbaselined ion engines for North-South Station-

4

keeping on it’s next generation ofcommunications satellites.16

Figure 4: Ion Engine

In determining the thruster power curve, Figure2, we took data from a wide range of ionengines.15,16,17,18,19,20,21,22 Figure 2 shows twodistinctly different trends for the ion enginesdepending on the diameter of the thruster.Thrusters larger than 20 cm follow a steeperpower versus specific impulse curve than thosesmaller than 20 cm. These smaller thrusters areintended primarily for station-keeping missions.Therefore, since this study is concerned withorbit transfer missions, data from thrusterssmaller than 20 cm is omitted.Ion engine efficiency was modeled using thesame form as the Hall thruster, and is shown inFigure 3.

One of the main concerns about both Hallthruster and ion engines for orbit transfer islifetime. Lifetimes of up to 7400 hours (308days) for the Hall thruster (SPT-100)6 and 4350hours (181 days) for the xenon ion engine21 havebeen demonstrated, with an 8000 hour (333days) test set to begin on the NSTAR ionengine.18 The primary concern is thrustererosion (wall erosion for the SPT, guard ringerosion for the ALT, and grid erosion for the ionengine) which will become even more of anissue with higher power thrusters.

Of paramount importance for any solar electricpropulsion vehicle is solar panel technology.

For orbit transfer missions, concentrator arraysoffer several advantages over conventionalplanar arrays.14 Concentrator arrays use opticsto focus solar radiation onto the solar cells, withconcentration ratios up to 100:1. These opticsprovide inherent radiation shielding that reducesthe need for a thick coverglass. This results inboth greater radiation resistance and reducedcell area compared to conventional planararrays. A concern of concentrator arrays is thatthey demand accurate 2-axis pointing,increasing the complexity of the orientationsystem.23 Pointing accuracies of ± 3° arerequired as compared to ± 18° for planararrays.24 Projections for concentrator arraysystems indicate a specific power approaching100 W/kg and a power to area ratio of well over200 W/m2, for technologies including galliumarsenide and multijunction arrays. Similarperformance can be achieved with planar arrays,but with much greater radiation degradation.

MODEL:

The centerpiece of the model is the payloadmass fraction equation based on the methoddeveloped by Messerole.3 This method isderived by starting with an initial massbreakdown:m m m m m m

m m m m mo pl pwr p tf ps

att adap ss cont rp

= + + + +

+ + + + +'

'

(2)

where the terms are defined in Table 1.

Term

Name Explanation

mo Initial Mass Total System mass atbeginning of E.P.transfer

mpl Payload Mass Useful on-station massmpwr Power Mass Mass of power

dependent components(thrusters, powerprocessors, solar arrays,and radiators)

mp PropellantMass

Mass of xenon used fortransfer

mtf Tank andFeedsystemMass

Mass of fuel tank andassociated components

mps’ Primary Mass of satellite’s

5

StructureMass

major structuralcomponents

matt AttitudeControl Mass

Mass of attitude controlsystem

madap Adapter Mass Mass of propulsion tosatellite adapter

mss’ SecondaryStructureMass

Mass of power relatedsystem structures

mcont ContingencyMass

Contingency mass forpower related systems

mrp RadiationProtectionMass

Mass of shielding toprotect payload fromradiation damage

Table 1: Mass Breakdown

The masses are divided by the total initialsystem mass to obtain mass fractions. Thepropellant tank and feedsystem mass fraction iscalculated by dividing by the mass of the fueland is expressed as the tank and feedsystemfraction, ftf. The primary structure mass,attitude control mass, and adapter mass arecombined into one term, the primary structurefraction, fps. The secondary structure andcontingency mass fractions are expressed interms of the propulsion system dry mass (powermass plus tank and feedsystem mass) and arecombined into the secondary structure fraction,fss. The radiation protection mass is expressedin terms of the payload mass using the radiationprotection fraction, frp. Combining these terms,Equation 2 can be rewritten to express thepayload fraction as:

( ) ( )m

m

f fm

mf

m

m

fpl

o

ps sspwr

ptf

p

o

rp

=

− − + + +

+

1 1 1

1 (3)

Other than the fractions, there are two terms inthe payload fraction equation. The first term isthe propellant mass fraction and the second isthe power mass to propellant mass ratio. Thepropellant mass fraction is calculated from therocket equation:25

m

mep

o

V g ITot o sp= − −1 ∆ (4)

Where ( )∆ ∆V f VTot perf= +1 and fperf is an orbit

transfer performance factor designed to accountfor thrust vector misalignment, off nominalthruster performance, and other contingencies.

The power mass to propellant mass ratio,(mpwr/mp), is calculated using the propellantmass flow rate to obtain:

m

m

m

t mpwr

p

pwr

t p

=&

(5)

Where tt is the thrusting time, equal to the triptime (t) multiplied by the non-occulated transitpercentage. It is taken to be 86%, which is thecase for a 180 day LEO to GEO mission, but isnot adjusted for changes in trip time or mission.Using the definition of specific impulse andefficiency, we can then rewrite Equation 5 as:

( )m

m

g I

t fpwr

p

o sp

t ppu t ave

=

2

2 η η α (6)

The term fave is the mission average powerfraction which accounts for solar arraydegradation due to radiation damage. It isdependent on the solar array technology and triptime. For planar arrays, it is 0.70 for a 180 daytransfer, for concentrator arrays, it is 0.97 forthe same trip time.3

The overall specific power (for the powerdependent components), α, is defined as:

α = =+ + +

P

m

P

m m m mB O

pwr

B O

a ppu t rad

. .L. . .L. (7)

Where PB.O.L. is the power at the beginning ofthe orbit transfer vehicle’s life. Then wedetermine the specific power by inverselysumming the component specific powers:

( )α α α α α= + + +− − − − −

a ppu t rad1 1 1 1 1

(8)

Solar array specific power, αa, is taken fromdata on concentrator array technologies to be100 W/kg.14

For the thruster specific power, αt, baselineconditions are taken from the SPT-100 and the

6

NSTAR ion engine. It is then possible to showthat if thruster mass is proportional to mass flowrate and mass flow rate is invariant with respectto input power, the specific power variesaccording to the relation:3

α αηη

αt too sp

spo

I

I= +

2

2* (9)

The constants αo and α* are determined by curvefitting actual thruster data as shown in Figure 5.For the ion engine, only large (diameter ≥ 20cm) ion engines are considered relevant. Theoptimal values found for these are summarizedin Table 2.

1500 2000 2500 3000 3500 4000 4500


0

100

200

300

400

500

600

700

800

900

1000

1100

Spec

ific

Pow

er (

W/k

g)

Hall Thruster: ModelIon Engine: ModelIon Engine: Actual

Figure 5: Thruster Specific Power versusSpecific Impulse

Thruster αo α* Ispo ηo

HallThruster

385.7 0.0 1600 0.49

IonEngine

347.0 -243.0 2500 0.62

Table 2: Thruster Specific Power Constants

For the power processing unit’s specific powerαppu, data was gathered from documented Hallthruster and ion engine PPUs.6,18 Additionally,due to a lack of other high power (>3 kW) spacequalified PPUs for xenon propellant systems,data from the US Air Force’s 26 kW arcjetElectric Propulsion Space Experiment’s (ESEX)PPU is included.26 The ESEX PPU is a 1991

design and is considered a conservative estimateof future PPU specific power performance. Theresulting specific power versus PPU input power(shown in Figure 6), was linearly curve fit withtwo straight lines. However, since power is anoutput quantity from this analysis, therecommended power per thruster is used todetermine the PPU specific power by assumingone PPU per thruster.

0 5000 10000 15000 20000 25000 30000

Power (W)

0

100

200

300

400

500

600

Spec

ific

Pow

er (

W/k

g)

Actual Specific Power (W/kg)Modeled Specific Power (W/kg)

ESEX

NSTAR

SPT-100

Figure 6: PPU Specific Power versusInput Power

Based on documented PPUs, the efficiency wasset to 93% for the Hall thruster6 and 90% for theion engine.17 The lower efficiency of the ionengine PPU is due to its more complex powerneeds.

Finally, for the radiator specific power, αrad, avalue of 32 W/kg is assumed for a system with aPPU of 92% efficiency.17

The required beginning of life input power for agiven thruster system can be determined fromthe overall specific power, the power mass topropellant mass ratio, the propellant massfraction, and the initial mass:

( ) ( )P mg I

t feB O L o

o sp

t pcu t ave

V g ITot o sp

. . ./= − −

2

21

η η∆ (10)

MISSIONS:

Four orbit raising missions are considered inthis paper:

1. Low earth orbit (LEO) to geosynchronousearth orbit (GEO) all-electric propulsiontransfer

7

2. LEO to LEO (higher) orbit transfer forconstellations of LEO satellites

3. Geostationary orbit insertion with partialchemical propulsion (orbit topping)

4. Reusable Orbit Transfer Vehicle (ROTV)for LEO to GEO transfer

These missions represent a cross section offuture orbit transfer missions for which electricpropulsion is a viable option. High Isp and ∆Vmissions involving interplanetary transit are notconsidered since they do not fall with the US AirForce’s current mission parameters.

Though not a direct part of the orbit transfermission, the launch vehicle makes a majorimpact on mission parameters, especially withregards to launch mass and payload faring size.The launch vehicles considered in this studywere the McDonnell Douglas Delta II (7920),the General Dynamics Atlas IIAS, and theLockheed Martin Titan IV with Hercules SolidRocket Motor Upgrade (SRMU). Launches toLEO were not to the traditionally used 100nautical mile (185 km), 28º circular orbit. Thiswas due to the fact that the large solar arraysnecessary for electric propulsion OTVs coupledwith their low thrust create a thrust to drag ratiodangerously close to one. It was found however,that by raising the launch altitude to 300 km,approximately 98% of the LEO mass could beretained while raising the thrust to drag ratio toalmost ten. In all cases, the largest payloadfaring available was used with the booster toinsure that maximum space was provided forlarge components such as solar arrays. Withthese considerations, the masses inserted intothe 300 km LEO orbits are summarized in Table3.

Booster Launch MassDelta II 4915 kgAtlas IIAS 8300 kgTitan IV 21150 kg

Table 3: LEO Masses27

Thruster sizing is a major concern of satellitemanufacturers. Due to the existence of aparametric series of Hall thrusters, it is possibleto derive a sizing relation with respect tospecific impulse based on the thrusters’characteristics. This relation is shown in Figure

7 with existing and planned thrusters identified.However, the physical dimensions of the ionengine are not as sensitive to changes in specificimpulse (and thus power) as the Hall thruster.Therefore, all ion thrusters are assumed to be 30cm in this study. Thirty centimeters is theaverage size of the large ion engines consideredin this study.

500 1000 1500 2000 2500 3000 3500


50

100

150

200

250

300

350

400

450

Thr

uste

r D

iam

eter

(m

m)

SPT-290

SPT-140

SPT-200

SPT-100

SPT-70

Figure 7: Hall Thruster Diameter versusSpecific Impulse

In this analysis, ∆V is calculated using the firstorder approximation:

∆V V V V Vo f o f= + −2 2 2 cosϕ (11)

The effect of gravity on ∆V, due to the longerelectric propulsion trip times is ignored in thisanalysis. In other cases, such as the orbittopping mission, ∆Vs had previously beencalculated using sophisticated orbit transfercodes like SECKSPOT28 which include thegravity ∆V. These previously calculated ∆V’swere manually entered for analysis. In no caseswere the two ∆V methods mixed.

The mission parameters (∆V and massfractions) for the four missions examined areshown in Table 4.

LEO-GEO Transfer:

The first mission examined was a basic LEO-GEO orbit transfer. This mission is applicablenot only to communications satellites, but also toproposed observation and reconnaissancemissions. Transit times of 180, 270, and 360days were examined for launches on all three

8

launch vehicles. The values used in the payloadfraction equations are shown in Table 4.

Key parameters determined are the payloadmass fraction, the array output power required,and the number of thrusters necessary. We notethat payload mass fraction is independent ofinitial mass and thus invariant with launch

vehicle. However, the power and number ofthrusters is strongly dependent on the initialmass launched into LEO.

It is clear that no single parameter determinesthe best thruster configuration for any givenmission.

Parameter LEO-GEO LEO Constellation Orbit Topping Reusable OTV∆V 5233 m/s 276.8 m/s Varies 5233 m/sfperf 0.01 0.01 0.01 0.01fps 0.06 0.06 0.06 0.20fss 0.10 0.10 0.10 0.10ftf 0.12 0.12 0.12 0.12frp 0.03 (t/180) 0.20 0.03 (t/180) 0.03 (t/180)fave 1-0.03 (t/180) 0.97 1-0.03 (t/180): t ≥ 180

0.97: t < 180favei

i

n

=∏

1

n = transfer numberfavei 1-0.03 (t/180)

Table 4: Mission Parameters

It must be a combination of critical factorsincluding trip time, payload mass fraction,power required, and number of thrusters.

Figures 8 through 10 show that for LEO-GEOtransfers Hall thrusters deliver a highermaximum payload fraction than ion engines.This is driven by the fact that ion enginesystems have a lower overall specific powercompared to Hall thruster systems. FromEquation 6, we see that this gives a higherpower system mass to propellant mass ratio,resulting in a lower payload fraction fromEquation 3. The major factor driving the overallspecific power differential is the lower specificpower of the ion engine itself. This is clearlyseen from Figure 5. A secondary effect is thelower power per thruster for the ion engine (asseen from Figure 2), resulting in ion enginePPUs having lower specific powers, from Figure6, for similar specific impulses. ComparingFigures 8 through 10, we note that increasingtrip time decreases the differential between Hallthrusters and ion engines. The decrease resultsfrom Equation 6, where increasing trip timelessens the overall effect of specific power onpayload fraction.

1000 1500 2000 2500 3000 3500 4000


0.05

0.1

0.15

0.2

0.25

0.3

0.35

Pay

load

Fra

ctio

n (-

)

Thruster:Hall ThrusterIon Engine

Figure 8: Payload Mass Fraction versus SpecificImpulse, Trip Time = 180 Days

1000 1500 2000 2500 3000 3500 4000


0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

Pay

load

Fra

ctio

n (-

)


Figure 9: Payload Mass Fraction versus SpecificImpulse, Trip Time = 270 Days

9

1000 1500 2000 2500 3000 3500 4000


0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

0.55P

aylo

ad F

ract

ion

(-)


Figure 10: Payload Mass Fraction versusSpecific Impulse, Trip Time = 360 Days

Figure 8 shows a peak in the payload fraction.This peak occurs for all trip times, but as triptime is increased, it occurs at progressivelyhigher specific impulses. The peak can beexplained as follows. As specific impulse isincreased, we see from Equation 4 that thepropellant fraction decreases. Looking at thepower mass/propellant mass term in Equation 6,we see that the numerator increases with thesquare of the specific impulse. The terms in thedenominator dependent on the specific impulse,the specific power and the thruster efficiency,also increase with specific impulse, but not asrapidly as the numerator (except at low specificimpulses, where the efficiency term as seen inFigure 3 increases rapidly). Thus, the powermass/propellant mass term increases withincreasing specific impulse. As specific impulseis increased, the increasing powermass/propellant mass term is countered by thedecreasing propellant mass fraction in Equation3, resulting in the initial rise in payload fractionversus specific impulse that we see in Figure 8.However, we note from Equation 4 that the rateof decrease of the propellant fraction slows withincreasing specific impulse. Therefore,eventually the increase in power mass/propellantmass will overcome decreases in propellantfraction, resulting in the eventual decrease inpayload fraction that is also observed in Figure8. To summarize, increases in power systemmass fraction with increasing specific impulsebegin to dominate the system, eliminating roomfor payload.

Next, we look at required power versus specificimpulse as shown in Figures 11 through 13.

1000 1500 2000 2500 3000 3500 4000


0

50

100

150

200

250

Arr

ay O

utpu

t P

ower

Req

uire

d (k

W)

Thruster/Booster:Hall Thruster: Delta IIHall Thruster: Atlas IIASHall Thruster: Titan IVIon Engine: Delta IIIon Engine: Atlas IIASIon Engine: Titan IV

Figure 11: Required Power versus SpecificImpulse: Trip Time = 180 Days

1000 1500 2000 2500 3000 3500 4000


0

20

40

60

80

100

120

140

160

Arr

ay O

utpu

t P

ower

Req

uire

d (k

W)



1000 1500 2000 2500 3000 3500 4000


0

20

40

60

80

100

120

140

Arr

ay O

utpu

t P

ower

Req

uire

d (k

W)



For a given specific impulse, trip time, andpayload the power levels are very similarbetween Hall thruster and ion engine systems, asexpected from Equation 10, since the only

10

differences are the thruster and PPU efficiencies.However, when comparing power levels forequal payload fractions, Hall thruster systemsrequire significantly lower power. Givenprojected trends in GEO satellite power levels, itappears from figures 11 through 13 thattransfers of Delta II and Atlas IIAS payloadswith trip times in the 270 to 360 day rangeshould be feasible within the next 15 years.However, it appears that transfers of Titan IVclass payloads will be impractical for some timeto come.

1000 1500 2000 2500 3000 3500 4000


0

5

10

15

20

25

Num

ber

of T

hrus

ters

(-)

Thruster/Booster:Hall Thruster: Delta IIHall Thruster: Atlas IIASHall Thruster: Titan IV

Ion Engine: Delta IIIon Engine: Atlas IIASIon Engine: Titan IV

Figure 14: Number of Thrusters versus SpecificImpulse: Trip Time = 180 Days

1000 1500 2000 2500 3000 3500 4000


0

5

10

15

20

25

Num

ber

of T

hrus

ters

(-)




1000 1500 2000 2500 3000 3500 4000


0

5

10

15

20

25

Num

ber

of T

hrus

ters

(-)




Next, we look at the number of thrusters,operating at nominal conditions, necessary toaccomplish the orbit transfer. From Figures 11through 13, we see that the total power for amission goes up with increasing specificimpulse. However, examining Figure 2, we seethat the power capacity per thruster alsoincreases, and at a higher rate. Therefore, thenumber of thrusters, determined by dividing thetotal power by the power per thruster, decreaseswith increasing specific impulse. Forredundancy purposes, we never consider fewerthan two thrusters.

From Figures 14 through 16, we note that atlower specific impulses an impractical numberof thrusters is required. As specific impulse isincreased, the number of Hall thrustersnecessary drops to realistic levels (less than 10)in all cases. Ion engines, however, only reachthe ten thruster cutoff for the 270 and 360 dayDelta II missions and the 360 day Atlas IIASmissions. The Titan IV mission does not appearin Figures 14 and 15 since the minimumnumber of ion engines is 50 and 32 respectively.Ion engine systems require higher power and,more importantly, ion engines increase in powercapacity per thruster more slowly than Hallthrusters, as can be seen from Figure 2.

Faring volumetric constraints of the propulsionsystem are anticipated to be another majorconcern of satellite manufacturers. LEO-GEOtransfers require a large number of presentgeneration thrusters. For example, a Delta IIlaunch with an 180 day trip time delivers

11

maximum payload using three 210 mm(approximate) Hall thrusters operating at 2500 sspecific impulse and 35.6 kW of total power.This results in a total thruster area ofapproximately 0.26 m2. By comparison,maximum payload for an ion engine system isobtained from eighteen thrusters (30 cmdiameter), operating at 3200 s specific impulseand 45.5 kW total power. This is 3.16 m2 ofthruster area, or over twelve times the area of aHall thruster system. A more practical systemwould probably have four larger ion thrusters, ofapproximately 64 cm each (the equivalent gridarea of the eighteen smaller thrusters), howeverthese larger thrusters may have significantlydifferent operating conditions.

It appears in general that Hall thrusters are thebest choice for LEO-GEO transfers. IgnoringTitan IV payloads, Hall thrusters in the specificimpulse range of 2000 to 3000 s can deliverhigh payload fractions at power levels that areconsistent with future trends for GEO satellites,and can do so with a realistic number ofthrusters.

The true benefit of electric propulsion becomeseven clearer when it is compared to chemicalpropulsion systems. Looking at the maximumpayload deliverable to GEO and payload fractionfor the Hall thruster cases examined and for allchemical systems in Table 5, we see largeincreases in payload capacity. This increase inpayload capacity gives three major benefits.First, and most obvious, it permits increases inuseful payload delivered to orbit. This increasedpayload could result in increased satellitehardware or additional propellant for station-keeping to lengthen the life of the satellite.Second, increases in payload capacity will relaxconstraints on satellite design, allowing greaterflexibility. Third, and perhaps mostimportantly, it allows for launch vehicledownsizing in some cases, as can be seen for270 and 360 day transfers for the Delta IIcompared to the chemical Atlas IIAS. Theobvious deterrent to these systems is the triptime required.

System Delta II AtlasIIAS

Titan IV

Chemical27 900 kg18.3%

2100 kg22.6%

4550 kg21.5%

180 Day EP33.3%

1637 kg 2766 kg 7047 kg

270 Day EP45.0%

2214 kg 3739 kg 9528 kg

360 Day EP51.0%

2506 kg 4232 kg 10784 kg

Table 5: Chemical and Electric PropulsionGEO Masses

LEO Constellations:

The second case is representative of transfersneeded for LEO constellations. The particularcase studied here is a space based observationsystem. It is based on the use of a largeconstellation (1148 satellites) in a Walker orbitat an altitude of 500 miles for theater levelreconnaissance as discussed by Fiedler andPreiss.29 Though this concept was notrecommended in their study due to the cost ofsuch a large number of satellites, the operatingparameters nevertheless provide an excellentreference point for other potential LEO satelliteconstellations. For this case, we use a ∆V of276.8 m/s, which is the ∆V from 162 to 500miles. This mission ∆V does not include anyinclination change or repositioning during thesatellite’s lifetime. Here, trip times of 180, 90,and 30 days are considered using each of thethree launch vehicles.

There are two other parameter changes for thismission. The first is that since these satelliteswill be operating in a much more radiationallyintense orbit than geosynchronous satellites, theradiation protection fraction, frp (the mass of theradiation shielding necessary to protect thepayload, expressed in terms of the mass of thepayload), is set to 0.20 as compared to 0.03 to0.06 for the LEO-GEO case. Second, theaverage power fraction is maintained at the 180day level, which is 0.97 for all missions. Theparameters used are summarized in Table 4.

In proceeding to the analysis of LEOconstellations, one notes that the same quantitiesare of interest as for the LEO-GEO transfercase.

12

1000 1500 2000 2500 3000 3500 4000


0.35

0.4

0.45

0.5

0.55

0.6

0.65

0.7

0.75

0.8

Pay

load

Fra

ctio

n (-

)



1000 1500 2000 2500 3000 3500 4000


0.3

0.35

0.4

0.45

0.5

0.55

0.6

0.65

0.7

0.75

Pay

load

Fra

ctio

n (-

)



1000 1500 2000 2500 3000 3500 4000


0.3

0.35

0.4

0.45

0.5

0.55

0.6

0.65

0.7

Pay

load

Fra

ctio

n (-

)



Looking at Figure 17, we see that for 180 daytrip times, Hall thrusters and ion engines arenearly equal with regard to payload fraction. Astrip time is decreases, we note from Figures 18and 19 that the Hall thruster delivers a higherpayload fraction. The reasons for this behavioris the same as in the LEO to GEO case. Thehigher payload fractions compared to Figure 8

result from reductions in propellant fraction atthe lower ∆V. For Hall thrusters of higher than1500 s specific impulse, payload fractionsgreater than 0.65 can be achieved at trip timesas short as 30 days, assuming adequate power isavailable.

Examining the power requirements for LEOconstellation transit shown in Figures 20through 22, we see that they are much lowerthan those for LEO-GEO transfers with thesame initial mass. However, given the lowerpower levels expected for LEO constellations,Atlas transfers of 30 days and Titan transfers of90 days appear to be at the upper boundary ofwhat is likely to be possible in the next 10 to 15years.

1000 1500 2000 2500 3000 3500 4000


0

2

4

6

8

10

12

14

Arr

ay O

utpu

t P

ower

Req

uire

d (k

W)



1000 1500 2000 2500 3000 3500 4000


0

5

10

15

20

25

30

Arr

ay O

utpu

t P

ower

Req

uire

d (k

W)



13

1000 1500 2000 2500 3000 3500 4000


10

20

30

40

50

60

70

80A

rray

Out

put

Pow

er R

equi

red

(kW

)



The number of thrusters reduces similarly to theLEO-GEO case, but due to the smaller amountof thrust necessary, there are fewer needed. Theonly case where a system with fewer than 10thrusters is not possible is the 30 day transfer ofa Titan IV class payload using ion engines.

Orbit Topping:

One of the most intriguing applications ofelectric propulsion is orbit topping. Proposed byFree30 and further studied by Oleson, et al.28 andSpitzer,31 the concept involves performing aninitial portion of a LEO-GEO transfer usingchemical propulsion, with the final GEOinsertion done using electric propulsion. Theprimary advantage of this approach is that itallows a significant propellant mass savingsover an all-chemical transfer without the longtrip times of an all-electric transfer, whichtranslates into a payload increase.

In the work of Oleson, et al., the SECKSPOTorbit transfer code was used to optimize alaunch using an Atlas IIAS with a Centaurimpulsive chemical stage. A combination of on-board chemical propulsion and electricpropulsion is then used to insert the satellite intoa geostationary orbit. In their study, the amountof on board chemical ∆V was decreasedincrementally while the amount of electricpropulsion ∆V was increased (the Centaur stageis the same in all cases). They consideredseveral thrusters (arcjets, Hall thrusters, and ionengines) and two power levels, and found thatthe ∆V’s for the transfer were approximately thesame for all cases. The study indicated that the

greatest mass gains could be made using an ionengine (2.5 kW NSTAR), with a stationaryplasma thruster (1.5 kW SPT-100) coming inclose behind. It was noted in Reference 28,however, that a more appropriately poweredSPT (2.5 kW) may have increased benefits.

In our study, thrusters of equal power levelswere compared against each other. Power levelsup to 5 kW per thruster were studied here,compared to 1.5 kW Hall thrusters and 2.5 kWion engines examined by Oleson, et al. Thehigher specific power devices investigated inthis paper illustrate even greater advantages forelectric propulsion orbit topping than thoseshown by Oleson, et al.

Using the electric propulsion starting conditionsfrom the work of Oleson, et al., side-by-sidecomparisons of Hall thrusters and ion enginesoperating at 1.66, 2.5, and 5.0 kW were madefor missions with 10 and 15 kW of total power.These total power levels determine the numberof thrusters used for each case. Operatingconditions were the same as the standard LEO-GEO transfer, with ∆V based on theSECKSPOT calculated orbits. The trip time andpayload are calculated by decreasing trip timeuntil the mission power level is reached. Arraydegradation was determined through the averagepower fraction, fave which was maintained at0.97 for trip times of under 180 days andincreased for trip times greater than 180 daysusing the same form as for LEO-GEO transfers.These fractions are summarized in Table 4.

0 50 100 150 200 250 300 350 400

Trip Time (days)

1600

1800

2000

2200

2400

2600

2800

3000

Fin

al M

ass

(kg)

Hall Thruster: 1.66 kWHall Thruster: 2.5 kWHall Thruster: 5 kWIon Engine: 1.66 kWIon Engine: 2.5 kWIon Engine: 5 kW

Figure 23: Final Mass versus Trip Time:Power = 10 kW

14

0 50 100 150 200 250

Trip Time (days)

1600

1800

2000

2200

2400

2600

2800

3000

Fin

al M

ass

(kg)

Hall Thruster: 1.66 kWHall Thruster: 2.5 kWHall Thruster: 5 kWIon Engine: 1.66 kWIon Engine: 2.5 kWIon Engine: 5 kW

Figure 24: Final Mass versus Trip Time:Power = 15 kW

Since power and number of thrusters are a givenfor orbit topping, the only pertinent results arepayload and cost. Data is presented in the sameway as Oleson, et al., with final mass plottedversus trip time in Figures 23 and 24. Finalmass is defined as useful on-orbit massincluding payload, power systems, structure, andattitude control. It also includes a mass penaltyfor array degradation (i.e., 1-fave multiplied bythe mass of the array). Each data pointrepresents an individual ∆V case.

The orbit topping scenario is power limited byfixing the amount of power available for thepropulsion system. This differs from the twoprevious cases which were trip-time limited. Byallowing trip time to vary, we see from Figures23 and 24 that for a given amount of ∆V, ionengines can deliver more payload from theintermediate orbit to GEO, however they requiremore time to do so. The additional payload andlonger trip time results from the fact that at agiven input power, the ion engine has a higherspecific impulse. From Equation 4 and otherbasic rocket equations, we see that the higherspecific impulse results in lower propellant mass(and thus higher payload mass) and longer triptimes. Additionally, in both Figures 23 and 24,there is a much larger spread between the powerlevels for the various ion engines than for theHall thrusters. This spread is due to the largerincrease in specific impulse in going from a 1.66kW ion engine to a 5 kW ion engine, comparedto the same change in power level for the Hallthrusters. Increasing power from 10 kW to 15kW does not significantly affect the final massdelivered, but it does reduce trip time byapproximately 33%.

We note that for trip times on the order ofsatellite check-out periods (~30 days) payloadgains of over 100 kg can be achieved. Forcomparison, Oleson, et al. calculated that asystem using all chemical propulsion for orbittransfer would have a mass of 1723 kg if it usedSPTs for North-South Station Keeping (NSSK)and 1748 kg using ion engines.

The choice of thruster type for this mission willdepend on the requirements of the user. Forminimum trip time on a given mission (∆Vcase), 1.66 kW SPTs or other low specificimpulse Hall thrusters would be the best choice.However, if the user wants to maximize payloadfor a given mission, 5 kW or higher power ionengines would be the best thruster.

Reusable Orbit Transfer Vehicles:

The final mission examined in this study is areusable orbit transfer vehicle (ROTV). Theparticular concept for ROTV studied in thispaper is a modification of that proposed as partof a recent US Air Force study.32 In thatconcept, there were two modules: a reusablepower system composed of the solar arrays, bus,and docking module; and an expendablepropulsion system, launched with the payload,comprised of thrusters (arcjets, resistojets, or ionengines), propellant tank, and power processingunit. In our study, the thrusters were changed toeither Hall thrusters or ion engines, and the PPUwas moved to the power system along with thethermal radiator.

For this study, it is assumed that the propulsionsystem/payload is launched on a Delta II, usingthe full payload capability. The power system isalso launched on a Delta II, but due to the massbreakdowns of the system, the payload capacityof the second Delta II will not necessarily beused by the power system (the unused payloadcapability can be used in some cases for asecond power system module, or for auxiliarypayloads). The actual mass launched on thesecond Delta II is determined as a function ofthe trip time and the specific impulse of thethruster used.

The thrusters examined are Hall thrusters withspecific impulses of 1600, 1900, 2200, and 2500

15

seconds and corresponding power levels fromFigure 2. The corresponding ion engine specificimpulses were determined by matching thethruster input powers to those of the Hallthrusters.

Missions where the power system is used for upto five complete transfers from LEO to GEO andback were studied. Initial trip times of 180, 270,and 360 days each way were examined.Subsequent transfers will take more time due tosolar array degradation.

The parameters for this mission are summarizedin Table 4. There are two differences betweenthe ROTV and standard LEO-GEO transfer.First, due to the requirement of on-orbitcoupling and decoupling of the power andpropulsion systems, the adapter mass must beincreased. There are three docking modulesnecessary in this concept: one to separate thepayload and one each on the power system andpropulsion system to dock the two together. It isassumed that each module comprises 5% of thetotal vehicle mass. This raises the primarystructure fraction to 0.20 as compared to 0.06for the other missions examined. Additionally,since the power system makes multiple tripsthrough the Van Allen belts, the arraydegradation is cumulative. Therefore, fave isdetermined by multiplying the individualtransfer power fractions (favei).

For the reusable orbit transfer scenario, theinitial masses were computed by determiningthe mass breakout of the system between theexpendable payload and propulsion system andthe reusable power system. The ratio ofexpendable component mass to reusablecomponent mass was lower for the ion enginethan for the Hall thruster. Since it was assumedthat the expendable/payload launcher is a fullcapacity Delta II, this results in a more massivereusable power system for the ion engine andthus a larger overall starting mass in LEO.

The mass transported to GEO includes both thesatellite payload and the propellant for thereturn trip. Satellite payload fractions areshown in Table 6 for the thruster configurationsstudied, both Hall thrusters and ion engines.These fractions represent the satellite payloadmass (not including the propellant for the return

trip), divided by the overall starting mass inLEO.

Cases with no payload listed are ones for whichthe payload capacity was insufficient totransport enough fuel to return the ROTV in thespecified trip time even with zero payload.

Thruster Initial TripTime (days)

Satellite MassPercentage

Hall Thrusters1600 s, 1.04 kW 180 ---

270 3.56 %360 10.5 %

1900 s, 2.45 kW 180 ---270 12.5 %360 18.7 %

2200 s, 5.08 kW 180 2.95 %270 16.6 %360 22.9 %

2500 s, 9.59 kW 180 5.09 %270 19.1 %360 25.7 %

Ion Engines1930 s, 1.04 kW 180 ---

270 ---360 1.22 %

3195 s, 2.45 kW 180 ---270 17.3 %360 25.7 %

4100 s, 5.08 kW 180 ---270 18.2 %360 27.5 %

5092 s, 9.59 kW 180 ---270 15.8 %360 26.5 %

Table 6: ROTV Delivered Satellite Masses

The other critical parameter is the powerrequired for these cases. Due to solar arraydegradation, the available power will be highestfor the first transit; subsequent trips will haveless power available. Since payload isconsidered constant, the trip time must thereforeincrease for subsequent round trips. The initialpower required is shown in Table 7 and the triptime increase is shown in Figure 25. (We notethat for any given round trip, the downward legcan be made in the same amount of time or lessthan the upward, since the mass transported hasbeen decreased by the mass of the payload andthe fuel used in the upward trip. Therefore, the

16

power required is lower. This analysis assumesthat the downward trip time is the same as theupward leg.)

Thruster Initial TripTime (days)

PowerRequired

Hall Thrusters1600 s, 1.04 kW 180 ---

270 27.8 kW360 19.7 kW

1900 s, 2.45 kW 180 ---270 29.1 kW360 20.7 kW

2200 s, 5.08 kW 180 55.3 kW270 31.7 kW360 22.4 kW

2500 s, 9.59 kW 180 61.8 kW270 34.8 kW360 24.5 kW

Ion Engines1930 s, 1.04 kW 180 ---

270 ---360 25.3 kW

3195 s, 2.45 kW 180 ---270 49.1 kW360 33.4 kW

4100 s, 5.08 kW 180 ---270 64.4 kW360 42.6 kW

5092 s, 9.59 kW 180 ---270 86.9 kW360 55.1 kW

Table 7: ROTV First Transit ArrayOutput Power

100 200 300 400 500 600 700 800 900 1000

Trip Time (days)

1

2

3

4

5

Tra

nsit

Num

ber

(-)


Figure 25: ROTV Trip Times withTransit Number

From Table 6 we see that for equivalent thrusterpower and trip time, the ion engine delivers ahigher satellite payload fraction than the Hallthrusters at high specific impulses. The overallmass fraction delivered to GEO (satellite plusreturn trip fuel) is lower for the ion engine, butsince it operates at higher specific impulses itrequires a lower propellant fraction for thereturn trip, allowing a higher satellite payloadfraction. In Table 7, we see that for equivalentcases the power required for the initial roundtrip is much higher for the ion engine than forthe Hall thruster, due to their higher specificimpulse and initial payload. Since the powerper thruster is fixed, this means that ion enginesystems will require more thrusters (assumingpresent day designs) than Hall thruster systems.Finally, from Figure 25 we see that the trip timeincrease for subsequent transfers is moresubstantial for the Hall thruster than for the ionengine.

Comparing the payload fraction data in Table 6to the standard LEO-GEO transfer as shown inFigure 10, payload capability is less for ROTV’sbecause of the need to carry propellant for thereturn trip. Figure 25 shows that in spite of theuse of highly resilient concentrator arrays, forinitial trip times greater than 180 days, arraydegradation quickly drives the trip time tounacceptable levels for round trip number twoand beyond.

CONCLUSIONS:

The analysis in this study shows that if the triptime for a mission is fixed, then Hall thrusterscan deliver higher payload fractions, due to theirhigher specific power. If, however, the powerfor a mission is fixed and trip time is allowed tovary, ion engines can deliver greater payloadsince they typically operate at higher specificimpulses.

Examining all four missions and taking intoaccount power and trip time requirements, themission that seems most practical at this timewould be orbit topping. Significant payloadgains could be made for 10 or 15 kW systemswith trip times of the order of 30 days, which isapproximately the on-orbit check-out time formost satellites. Also practical in the near termwould be small orbit transfers for LEO

17

constellations, since they offer high payloadfractions with short trip times.

A full, all electric propulsion, LEO-GEOtransfer is practical using today’s technology, aslong as the user is willing to accept the longtransfer times. However, by performing theshorter term orbit topping and LEOconstellation type missions, the user can beintroduced to the benefits of electric propulsiontransfers without immediately suffering thelarge trip time penalties.

The same comments apply to an even greaterextent to reusable orbit transfer vehicles.Though there can be payload benefits comparedto chemical systems, the trip time penalties aremuch more severe than expendable LEO-GEOsystems. Other issues to consider for ROTVsinclude: increased power requirement andlogistical concerns (autonomous control isprobably necessary for a practical system).These combine to make ROTVs the most longterm of the missions examined here.

For all of these missions, there are concernswith regards to thruster size and payload faringvolume. This is especially true for ion enginesystems since they require a higher number ofthrusters and ion engines are typically physicallylarger than Hall thrusters.

It is important to note that all of the analysis inthe paper is based on the current generation ofion engines and Hall thrusters. Future thrustersthat operate at higher specific powers andthruster power to specific impulse ratios, couldincrease system performance. Work has alreadybeen done in this area, such as the 50 cm andlarger ion engines that have been tested byNASA33,34 and Hall thruster designs such as theSPT-N series.11 Advanced thrusters such asthese also offer other improvements includinghigher efficiency and lower beam divergence.

Improvements in other areas can also improveelectric propulsion capability. Solar arraytechnology continues to improve, with increasesin specific power and material density that willmake performance improvements for both Hallthrusters and ion engine systems possible.Improvements in the cryogenic storage of xenoncan benefit both types of thruster by reducing thetankage fraction.

ACKNOWLEDGEMENTS:

The authors would like to thank Mr. CraigClaus of Atlantic Research Corporation and Mr.James Sovey of NASA Lewis Research Centerfor their information about stationary plasmathruster and ion engine developmentrespectively; Dr. J. Messerole from BoeingDefense and Space Group for his assistance withthe analytical engine; and the other engineersand scientists of the Phillips Lab ElectricPropulsion Laboratory for their input andsuggestions.

REFERENCES:

1. Operational Effectiveness Cost Study (OECS)Final Report, Office of Aerospace Studies,December 1995.2. Grobman, F., PL/VT, PersonalCommunication, June 1996.3. Messerole, J.S., “Launch Costs to GEO UsingSolar-Powered Orbit Transfer Vehicles,” AIAAPaper 93-2219, June 1993.4. Kaufman, H.R., “Technology of Closed-DriftThrusters,” AIAA Journal, Vol. 23, No. 1, 1985,78-86.5. Bober, A.S., et al., “State of Work onElectrical Thrusters in the USSR,” IEPC Paper91-003, October, 1991.6. Day, M., et al., “SPT-100 SubsystemQualification Status,” AIAA Paper 95-2666,July 1995.7. Sankovic, J., Hamley, J., Haag, T.,“Performance Evaluation of the Russian SPT-100 Thruster at NASA LeRC,” IEPC Paper 93-094, September 1993.8. Garner, C.E., et al., “Cyclic Endurance Testof a SPT-100 Stationary Plasma Thruster,”AIAA Paper 94-2856, June 1994.9. Russian Electric Propulsion Seminar,Massachusetts Institute of Technology, 1991.10. Day, M., Rogers, W., Maslinekov, N., “SPT-100 Subsystem Development Status and Plan,”AIAA Paper 94-2853, June 1994.11. Clauss, C., “High-Power Stationary PlasmaThruster Propulsion Systems,” AtlanticResearch Corporation Memorandum, October,1995.12. Brophy, J.R., “Ion Thruster PerformanceModel,” NASA CR-174810, December 1984.

18

13. “Ion Thruster Operation,” NASA Spacelink -An Electronic Information System forEducators, http://spacelink.msfc.nasa.gov/NASA.Projects/ NASA.Launch.Vehicles/Proposed.Systems/ Ion.Propulsion.Engine/Ion.Thruster.Operation, April 1996.14. Pollard, J.E., et al., “Electric PropulsionFlight Experience and Technology Readiness,”AIAA Paper 93-2221, June 1993.15. Patterson, M.J., and Foster, J.E.“Performance of a ‘Derated’ Ion Thruster forAuxiliary Performance,” AIAA Paper 91-2350,June 1991.16. Beattie, et al., “Flight Qualification of an 18-mN Xenon Ion Thruster,” IEPC Paper 93-106,September 1993.17. Rawlin, V.K., and Majcher, G.A., “MassComparisons of Electric Propulsion Systems forNSSK of Geosynchronous Spacecraft,” AIAAPaper 91-2347, June 1991.18. Marcucci, M., “Overview of the NSTARProgram,” 7th Advanced Space PropulsionWorkshop, April 9, 1996.19. Ashley, S., “Electric Rockets Get a Boost,”Mechanical Engineering, December 1995, 61-65.20. Beattie, J.R., et al., “Xenon Ion PropulsionSubsystem,” Journal of Propulsion and Power,Vol. 5, No. 4, 1989, 438-44421. Beattie, J.R., et al., “Status of Xenon IonPropulsion Technology,” Journal of Propulsionand Power, Vol. 6, No. 2, 1990, 145-150.22. Groh, K.H., and Loebt, H.W., “State-of-the-Art of Radio-Frequency Ion Thrusters,” Journalof Propulsion and Power, Vol. 7, No. 4, 1991,573-579.23. Gledhill, K., Marvin, D., “Future Trends inSpace Photovoltaics,” AIAA Paper 94-3970.24. Pollard, J.E., and Janson, S.W., “SpacecraftElectric Propulsion Applications,” AerospaceReport Number ATR-96(8201)-1, TheAerospace Corporation, February 1, 1996.25. Jahn, R.G., Physics of Electric Propulsion,McGraw-Hill Book Company, New York, 1968.26. LeDuc, J.R., ESEX Project Engineer,Personal Communication, April 1996.27. Isakowitz, S.J., International ReferenceGuide to Space Launch Systems, SecondEdition, AIAA Publication, 1991.28. Oleson, S.R., et al., “Advanced Propulsionfor Geostationary Orbit Insertion and North-

South Station Keeping,” NASA TM-107018,July 1995.29. Fiedler, S., and Preiss, B., “GeosynchronousSpace Based Radar Concept Development forTheater Surveillance,” IEEE AerospaceApplications Conference Proceedings, Volume4, 1996, 77-90.30. Free, B., “High Altitude Orbit Raising withOn-Board Electric Power,” IEPC Paper 93-205,September 1993.31. Spitzer, A., “Novel Orbit Raising StrategyMakes Low Thrust Commercially Viable,” IEPCPaper 95-212, September 1995.32. McFall, K.A., Electric PropulsionLaboratory Research Engineer, PersonalCommunication, March 1996.33. Rawlin, V.K., “Performance of Large AreaXenon Ion Thrusters for Orbit TransferMissions,” NASA TM-102049, May 1989.34. Sovey, J.S., et al., “Ion ThrusterDevelopment at NASA Lewis Research Center,”NASA TM-105983, January 1993.

ANALYSIS OF HALL-EFFECT THRUSTERS AND ION ENGINES …

Documents

Transcript of ANALYSIS OF HALL-EFFECT THRUSTERS AND ION ENGINES …