II-

NASA Technical Memorandum 1031.9_0............

AIAA-90-2579 ..................

Arcjet Load characteristics _ _

John A 5Hamley .............................Lewis Research Center

Cleveland, Ohio _

, T

(NA_A-TM-103190) ARC JET LOAD

CHAPACTFRISTIC S (NASA) 18 p CSCL 2iH

N90-25181

G3/eo

Prepared for the

........... _21stlnternational Electric_ Propulsion Confer eence ___:-_ _

, _ . _cosponsored by the AIAA, DGLR, and JSASS

Orlando, Florida, July 18-20, 19901_IIL _ 7 -

https://ntrs.nasa.gov/search.jsp?R=19900015865 2018-06-29T15:42:03+00:00Z

ARC JET LOAD CHARACTERISTICS

John A. HamleyNational Aeronautics and Space Administration

Lewis Research Center

Cleveland, Ohio 44135

ABSTRACT

Experiments were conducted to define theinterface characteristics and constraints of 1 kW class

arcjets run on simulated decomposition products ofhydrazine and power processors. The impacts ofpower supply output current ripple on arc jetperformance were assessed by variation of the ripplefrequency from 100 Hz to 100 kHz with 10% peak-to-peak ripple amplitude at 1.2 kW. Ripple had nosignificant effects on thrust, specific impulse orefficiency. The impact of output ripple on thrusterlifetime was not assessed.

The static and dynamic impedances of the arcjetwere quantified with two thrusters of nearly identical

configuration. Superposition of an AC componenton the DC arc current was used to characterize the

dynamic impedance as a function of flow rate andDC current level. A mathematical model wasformulated from these data. Both the static and

dynamic impedance magnitude were found to bedependent on mass flow rate. The amplitude of theAC component was found to have little effect on thedynamic impedance. Reducing the DC level from 10to 8 amps led to a large change in the magnitude ofthe dynamic impedance with no observable phasechange. The impedance data compared favorablybetween the two thrusters.

Ap_pACC

I3(:f

iac

iarc

JKi

m

n

Pi

Prms

qi

NOMENCLATURE

Amperes peak-to-peak, A

Alternating currentArc jet dynamic impedance gainconstantDirect current

Frequency, HzAC arc current, A

Arc current (total), A

DC arc current, A

(:i-Gain constant for a specific

pole-zero pairTotal number of zeros

Total number of poles

Dynamic impedance pole

RMS power, W

Dynamic impedance zero

s Complex frequencyvariable, j2_xf

T Thrust,g

vac AC arc voltage, V

Varc Arc voltage (total),V

Vdc DC arc voltage, VZd Dynamic impedance,

0 Impedance phase angle,degrees

x Transportation lag, seconds

INTRODUCTION

Arcjets were first considered as advancedpropulsion devices in the 1950's, and this earlyresearch program carried into the 1960's. At thistime, NASA's interest was primarily in planetarymissions and research centered mainly on high

powered arcjets with hydrogen propellant. 1-4Ammonia was also considered as an alternate

propellant. Arcjets in the 1 to 2 kW class wereconsidered for an auxiliary propulsion role, and thesedevices were run with some success on hydrogen

propellant. 5,6 Limited efforts to run the devices atthese lower power levels on storable propellants metwith little success. 7 In 1965, a review of these

programs was presented by Wallner and Csika. 8Interest in arcjets renewed in the last decade,

especially low power devices run on storablepropellants. A 1-2 kW device was selected for north-south stationkeeping missions (NSSK) ongeosynchronous spacecraft. Current programs havedemonstrated stable operation of these devices over awide range of power levels and mass flow rates.Specific impulse values in the range of 450 to 550seconds have been demonstrated at power levels and

propellant pressures expected in a NSSK mission. 914

In response to the need for system hardware, pulsewidth modulated (PWM) switching power supplies

with integral starting circuits were designed toprovide reliable start-up and steady-state operationwith vortex flow stabilized arcjets 15. These powerprocessing units produced 13 kHz ripple ofapproximately 1.5 ampere peak-to-peak amplitude,however, the effect of ripple on arc jet performancewas unknown. Preliminary testing indicated arelatively fast response in the arc column, leading toa step response specification on the order of 500

Copyright © 1990 by the American Institute of Aeronauticsand Astronautics, Inc. No copyright is asserted in the

United States under Title 17, U.S. Code. The U.S. Govern-ment has a royalty-free license to exercise all rights underthe copyright claimed herein for Governmental purposes.

All other rights are reserved by the copyright owner.

l.ts.15-18No detailedwork to characterizethedynamicimpedanceofthearcjetwasdoneduringthedevelopmentof thepowerprocessors,however,apreliminaryinvestigationtocharacterizethedynamicimpedancewasconductedafterward,whichdevelopedthemethodsusedhere.19

Thispaperpresentstheresultsof initialeffortstoevaluateandcharacterizethearcjet/powerprocessorinterface.First,thestaticimpedanceof thearcjetwasquantifiedasa functionof DCinputcurrentoverarangeof 8to 11.5A.Theeffectof flowratechangeswerealsoinvestigatedoverthisrangeofinputcurrents.Thedynamicimpedanceof thearcwasquantifiedby injectinganACcomponentontotheDCarccurrentoverthefrequencyrangeof 100Hzto 100kHz.Effectsof variationsin flow rate,DC currentlevel,andrippleamplitudeon thedynamicimpedancewerealsoquantified.Theimpactsof rippleon thrusterperformancewereassessedwith10%peak-to-peakrippleamplitudebysimultaneousmeasurementof thrustanddissipatedpowerwhilevaryingtheripplefrequencyovertherangeof 100Hz to 100kHz. Thrust,specificimpulseandefficiencywerethenplottedasfunctionsof ripple frequency.Theissuesof lifetimeandelectromagneticcompatibilityasafunctionofrippleamplitudeandfrequencywerenotaddressedhere.

Thiseffortwill provideinputsfor criteriatodeterminetherippleacceptablein futurepowerprocessordesigns,whichcouldofferreductionsinmassinoutputfilters.Thedynamicimpedancedataarealsoapplicableto powerprocessorcontrolsdesignin termsof stepresponseandcompensation.Thiswillprovidemoredetailedinformationastothecharacteristicsofthearcjetandpowerprocessorasasystemtothepowerprocessordesigner.

APPARATUS

Arciet Thruster

Two modular thrusters of nearly identicalconfiguration were used. A cutaway schematic of thethruster is shown in Figure 1. The thruster has beendescribed in detail elsewhere. 20 The constrictor was

0.51 mm in diameter and 0.25 mm in length. A3.18 mm 2% thoriated tungsten welding rod with a

conical tip and 30 ° half angle was used for thecathode. The anode also was made from the same

material and had a 30° half angle convergent section

and a 20 ° half angle divergent section. The arc gapwas set to 0.58 mm.

Propellant Managcm¢nl

A 2:1 stoichiometric ratio of hydrogen andnitrogen gas was used as a propellant throughout thetesting which simulated the decomposition products

of hydrazine monopropellant. The flow rates weremaintained using commercially available mass flowcontrollers.

A block diagram of the power supply used ispresented in Figure 2a. A 150V 15A laboratorysupply with a series pass regulator to maintainconstant current was used to supply power for theexperiment instead of a standard arcjet powerprocessor. A diode in series with the regulatorprotected it from the high voltage pulse used to startthe arcjet` A shunt modulator injected an AC signalof variable frequency and amplitude onto the theregulated DC supply current. Figure 2b details theoperation of the modulator. The output of the

current regulator was represented as ireg and thecurrent through the modulator imod, was controlled

by the function generator. The arc current iarc wasexpressed as:

i_= i_g- imod (1)

where imod depended on the input from the function

generator. The modulator required a 2A DC biascurrent to operate in a linear fashion, thus, the DCcomponent of the arc current, idc was lower than theoutput of the current regulator. This bias currentremained constant while the device is in operation. Ifa sinusoid was input to the modulator the arc currentwas represented as:

i_ = idc + iacsin (2rift) (2)

where idc represents the difference between the

regulator current and the DC bias current of themodulator. The amplitude and frequency of iac were

dependent upon the input from the functiongenerator. This approach was used because thefrequency response of the current regulator wasinsufficient to modulate its output directly. Avacuum relay protected the modulator from the highvoltage ignition pulse. After the arcjet was started,the vacuum relay was closed and the DC bias on the

modulator was applied slowly, avoiding any abruptchanges in arc current.

The arcjet was started with a pulse transformerderived from the 1 kW arcjet power electronics. Itsoperation is described in detail elsewhere 15,18. The

pulse was initiated using a pushbutton on the frontpanel of the equipment rack. An interlock inhibited

the operation of the starting circuit if the vacuumrelay in the modulator were not open. Coaxial cablewas used to supply the current to the arcjet.

_saungamtmnArc current was measured using a non-intrusive

current probe and a current shunt. The current probewas a commercially available probe with abandwidth of DC to 5 MHz which measured both

DCandAC currentcomponents,whiletheshuntwasusedto measureDCcurrentonly.Arcvoltagewasmeasuredusinganisolatedoscilloscopeprobe,andstandarddigitalmultimeters(DMMs).Thesedeviceswerecapableof measuringtheACandDCcomponentsofthevoltageandcurrent.Thecurrentprobeandshuntwerelocatedatthefeedthroughtothe vacuumchamberandthearc voltagewasmeasuredusingpotentialleadsmountedto thethrustertominimizetheeffectsof thepowercableandconnectionimpedances.

A gain/phasemeterwasusedto measuretherelativemagnitudesandphasesoftheACarccurrentandvoltage.Therearetwo inputson themeter.ChannelA isusedasareferenceandtheratioof theACamplitudesofthetwochannelsaredisplayedas20log(B/A)alongwiththephaseanglebetweenthetwochannelsindegrees.Anoscilloscopewasusedto verify the magnitudesof the arc currentcomponentsandthefrequencyoftheACcomponentwasdisplayedonafrequencycounter.

Thrustmeasurementsweremadeusinga flexuredisplacementthruststandwhichwascalibratedinsitu. This thruststandis describedin detailelsewhere.17

Facilities

The experiments were carried out in twofacilities 20. All impedance measurements were done

with the thruster mounted in a small bell jar 0.64 min length and 0.64 m in diameter. A singlemechanical roughing pump provided 21,000 lpmpumping capacity. Pressures on the order of 100 Pawere maintained at the maximum propellant flow.

All performance measurements were carried out in avacuum tank 5 m in length and 1.5 m in diameter.Four 0.81 m diameter oil diffusion pumps backed bya lobe type blower and two mechanical roughingpumps maintained the background pressure at

approximately 0.05 Pa at maximum propellantflow.

PROCEDURE

To quantify the effect of power processor rippleamplitude and frequency on thruster performance itwas first necessary to establish a baselineperformance level with a DC input current. Thethrust level, specific impulse, and efficiency weredependent on the power input to the device, whichwas determined by the static impedance only for the

DC case. As ripple was introduced, the dynamic orAC impedance of the device was included in theequation used to determine the dissipated power.This relation is discussed in depth later. Because ofthis, the static and dynamic impedances of the arcjetwere quantified prior to performance testing.Performance data were taken on only the first of the

two arcjets, hereafter referred to as thruster 1. Thesecond arcjet was referred to as thruster 2. Dynamicimpedance data were taken for both arc jets.

All impedance data were taken in the smaller ofthe two facilities described previously. Staticimpedance data were taken using thruster 1 only.The static impedance was measured using a shuntresistor and a DMM to monitor the arc current andanother DMM to monitor the arc voltage. Potential

taps were added to the thruster and brought outthrough feedthroughs. The shunt resistor wasmounted at the power feedthrough. Both DMMswere battery operated to avoid ground loops.

Flow rates of 4.3x10 -5, 4.5x10 -5 and 4.8x10 -5

kg/sec were used to investigate the dependence of thedynamic impedance on flow rate. These flow ratesare representative of those expected on a NSSKmission. The flow rate to the thruster was set and

maintained constant throughout a test. The arc jetwas started and allowed to stabilize thermally at 10A

DC operating current and 4.3x10 -5 kg/sec flow for30 minutes. The current was then reduced to 8A and

a 10 minute equilibration period followed. Staticvoltage and current data were then taken. Thisprocess was repeated increasing the current 0.5A DCincrements for each data point. Ten minutes wereallowed at each setpoint for equilibration.Theprocess continued to the 11.5A DC level and wasrepeated at each flow rate.

To quantify the dynamic impedance it wasnecessary to inject an AC component onto the DCarc current. This AC arc current caused an AC

component to appear on the arc voltage. Thedynamic impedance of the arcjet was defined as theratio of this AC arc voltage to the AC arc current. Amagnitude and phase were associated with this

impedance, and in these data the arc current was usedas the reference for the phase angle. The magnitude

of the impedance in dB was _given by:

l zal = 201ogt0[v'c/ [dB] (3)

and the phase angle 0 was expressed as:

These data can then be presented in the form ofBode plots and a mathematical description of thefollowing form:

11

I'l( s + qi)i=l

Za(s ) = K i (5)m

FI (s+Pi)i=l

Can be obtained empirically. The zeros of thecharacteristic equation, qi, are roots of thenumerator, and the poles, Pi are the roots of the

denominator. It has been shown that the presence of

apoleresultsinadecreasein themagnitudeof thecharacteristicequationof 20dBperdecadeat thecomerfrequencyassociatedwiththepole,andthephaseanglecontributionis-90°. A zerowill causethemagnitudetoincreaseata rateof 20dBperdecadeatthecomerfrequencywitha90° phaseanglecontribution.Thecomerfrequenciesaredeterminedbythenumericalvalueof thepolesandzeros.Thetheoryof systemfunctionsandtheirpolesandzerosisexplainedindetailelsewhere.21Theintentof theof theexperimentswastomapthepolesandzerosifpresentinthefrequencyrangetested.

Thruster1wasinstalledinthebelljarasdescribedforthestatictests.AcurrentprobewasaddedatthefeedthroughtomonitortheinjectedACarccurrentcomponent,andanisolatedoscilloscopeprobewasusedtomonitorthecorrespondingACarcvoltage.ThemassflowratewasadjustedtothedesiredlevelandthearcjetstartedwiththeDCarccurrentlevelsetto 10A.ThisDCcurrentlevelwasmaintainedthroughoutthetestingof thruster1.A 30minuteequilibrationperiodfollowed.Theoutputof thecurrentregulatorwasincreasedto 12ADCandthemodulatorstarted.ThemodulatorDCbiascurrentwasadjustedto2A,reducingthearccurrentto10A.Thefunctiongeneratorwasadjustedtoafrequencyof100Hzandtheamplitudesetsothatthemodulatorinjecteda 1Ap_psinusoidalcurrentcomponent,anAC current level was that was maintainedthroughouteachtest.

TheoutputofthecurrentprobeamplifierwassetforACcouplingandinputto thegainphasemeterchannelA, theoscilloscope,andthefrequencycounter.Theoutputof thevoltageprobeamplifierwasinputtochannelBof thegain/phasemeterandtheoscilloscope.Thegain/phasemeterwassettodisplaytheratiooftheamplitudesB/Aandthephaseangleof channelB withrespecttoA.Thegainsoftheinstrumentationamplifiersweresetsuchthattheratiodisplayedwas20 lOgl0(Vac/iac).Withthisconfigurationthemagnitudeandphaseof thedynamicimpedancewerereaddirectly.Thefrequencywasthenvariedinstepsof oneoctavewithineachdecadefrom100Hzto 100kHz,verifyingtheripplefrequencywiththefrequencycounter.A five minuteequilibration period was allowed following eachchange in frequency in early tests but was found tobe unnecessary.

The procedure was repeated at all flow rates. Toquantify the repeatability of the data this test matrix

was repeated three times. Attempts to obtain data athigher frequencies were made but the data were found

to be unreliable due to the inductance of the powerleads in the bell jar.

This procedure was repeated for thruster 2 with thefollowing additions. The DC current level was

reduced to 8 ADC and data were taken at 10% rippleamplitude for the three flow rates to determine the

effect of DC current level on the dynamicimpedance. Three ripple amplitudes were also used atthe 10 ADC level to investigate the linearity of theimpedance. The ripple was varied in discrete stepsfrom 8 to 12%.

For performance measurements thruster 1 wasinstalled on a flexure thrust stand in the largervacuum facility. Calibration and operation of thethrust stand are outlined in detail elsewhere. 17 To

measure the arc current, the current probe wasinstalled near the vacuum feedthrough along with astandard 100mV/10A shunt. The arc voltage in thiscase was measured at the vacuum feedthrough usinga DMM and the isolated oscilloscope probe. The DCcurrent level was maintained at 10 ADC throughoutthe performance tests, and the ripple amplitude wasset at 10%. The arcjet was started at the lower flowrate and allowed to run for 30 minutes at the 10ADC current level. Thrust measurements were taken

at the DC operating point and used as a baseline forcomparison to the measurements with ripple. Theprocedure from this point is identical to that of thedynamic impedance measurements, and these datawere taken in tandem with the thrust measurements.

To ensure the accuracy of the impedance data, theinstrumentation was calibrated using a 50£1 RF loadresistor prior to the tests. The gain of each of theinstrumentation amplifiers was set so that thegain/phase meter read the impedance magnitudedirectly in dB. The impedance of the resistor wasmeasured to be 50.5_. It was assumed that the

imaginary component of the load impedance wasnegligible in the frequency range of interest,therefore, the dynamic impedance of the deviceshould have a magnitude of 34.1 dB (201og(50.5))and a phase angle of 0 degrees from 100Hz to100kHz. Since the device is purely resistive, the

dynamic and static impedances are identical. Anydeviations from the 50.5 ohm impedance were

assumed to result from gain errors and phase lags inthe amplifiers. The resulting errors were small, butall data taken using the thruster were correctedaccordingly.

RESULTS AND DISCUSSION

For a given configuration the performance of thearc jet was dependent on the mass flow rate and thepower dissipated in the device. To properly computedissipated power, it was necessary to quantify thearcjet impedance. For this reason, the impedance dataare presented in'st, followed by the performance data.

, ,tiOmmaaa The static impedance data for thruster 1 are

presented in the form of volt-ampere curves inFigure 3. The arc voltage decreases with increasingcurrent, and this behavior is typical for high pressure

4

arcs as described in detail elsewhere. 22 An increasein the mass flow rate resulted in an increase in thestatic impedance. This was due to the associatedincrease in arc chamber pressure and a reduction inthe mean free path between propellant molecules.The shortened free path length resulted in an increasein the collision rate between particles, and moreenergy was needed to maintain the necessaryionization levels at a given current. Thus, a largerelectric field was required, increasing the arc voltage.This phenomenon is described in detail elsewhere. 23

Dvni_nic Im_texlanceMagnitudeThe magnitude of the dynamic impedance for

thruster 1 is presented in Figure 4a, and the phasedata in Figure 4b. The DC arc current was 10A andthe AC component had an amplitude of 1Ap_p. Fora given flow rate it can be seen that the magnitudeof the impedance remains relatively constant from100Hz to 10kHz, with a 2 dB decrease in magnitudefrom 10kHz to 100kHz. There was some indication

of a possible minima in the impedance magnitude at80 kHz. In general, as the flow rate increased themagnitude of the dynamic impedance increased. Thisincrease was not linear over the frequency rangetested, but varied from approximately 0.2 to 0.5 dBfrom the 4.3x10 5 to the 4.8x10 -5 kg/sec mass flowrate.

Dynamic Im_t_dance Phase Angle

The phase angle of the impedance was 190° at100 Hz, and remained constant, within 10°, to 10kHz, where the phase angle began to decrease. The

phase angle remained greater than 90° for the rangeof frequencies tested, therefore, the dynamicimpedance had a negative real component for thisfrequency range. This indicated that the V/Icharacteristic for this frequency range maintained anegative slope. From these data it appears that theflow rate had little effect on the phase angle of theimpedance. A discontinuity in the magnitude andphase data occurred at 1 kHz. This was anexperimental artifact traceable to the input filters onthe gain phase meter. A 1 kHz low pass filter wasused for frequencies below 1 kHz. For higherfrequencies a 100 kHz filter was used. This filter wasnot responsible for the minima in the impedancemagnitude at 80 kI-Iz, as the tests on the RF loadverified flat response in this region.

Data Re_oeatabilityAs shown in Figures 5a and 5b, the dynamic

impedance tests on tests on thruster 2 yielded resultssimilar to those obtained with thruster 1. Themagnitude of the dynamic impedance of the twothrusters agreed to within 1 dB and the phase angleto within 5 degrees.

Mathematical ModelFrom these data, no poles or zeros can be mapped

because the magnitude of the impedance remainednearly constant. However, the phase angle of thedynamic impedance decreased as the ripple frequencyincreased. This phenomenon can be characterized bya transportation lag, which is mathematicallydescribed as follows:

-|'[

zd(s) = e (6)where s = j2rff, and x is the transportation lag inseconds. This function has a magnitude of 1because"

I cos ( 2nfx ) -j sin ( 2nfx ) I = 1 (7)The phase angle is 0 for low values of f andbecomes more negative as f increases 23. Themagnitude of the dynamic impedance was greaterthan 1 however, and this can be accounted for byintroducing a gain constant C. The absolute value ofthe gain constant can be obtained from the dynamicimpedance data for a given flow rate. The magnitudeof the dynamic impedance was found using equation3, thus, the value of the gain constant is:

Izdl [dg ]

C = 10 • (8)Where I Zd I was taken from the data of Figure 4a.

The phase angle of the arc impedance was 180 ° forlow frequencies and this can be accounted for bymultiplying equation 6 by -1, which introduces a180 degree phase shift for low frequencies. Thephase angle will decrease as frequency increasesdepending on the constant x in equation 6 can beestimated from the phase data of Figure 4b and wasfound to be on the order of 1.5x10 "6 seconds. Since

the phase data showed no significant dependence onflow rate, this value of x can be used for all flowrates tested. The gain constant C, however, isdependent on flow rate. Typical values were on theorder of 3.76 for the 4.8x10 -5 kg/sec flow. Thedynamic impedance can therefore be approximatelyexpressed as

-(15 xl06) s

Zd(S) =-C e (9)

for the frequency range of 100 Hz to 100kHz.Figures 6a and 6b compare the measured dynamicimpedance magnitude and phase with those predictedby equation 9. The data compares favorably with thecalculated values. The maximum error of 2 dBoccurs at 100 kHz. These values for the dynamicimpedance magnitude were higher than thosepredicted by the static V/I curves, which were on theorder of 6dB. This value was obtained in thefollowing manner. When a 1A peak-to-peak ACcurrent component was superimposed on the DC arccurrent, the arc current had a maximum value of10.5A and a minimum value of 9.5A. The differencebetween the corresponding arc voltages for these

current values quantified the dynamic impedancesince Z=V/I. The magnitudes obtained in the

frequency range tested were greater than this expectedvalue. This may indicate the presence of a pole-zeropair at very low frequencies. This is further

supported by the 190 ° phase shift at 100 Hz. A

phase angle of only 180 ° is expected from the static

curves. Further work is necessary to investigate thisphenomenon. From these data, there is no indication

that the regulator step response need be changed.

Thruster 2 was used to investigate the linearity ofthe dynamic impedance. If the dynamic impedancewas linear, the impedance magnitude should remain

constant regardless of the magnitude of the ACcomponent. Figure 7a presents the magnitude dataand Figure 7b the phase data for a DC current of10A and a fixed flow rate. The magnitude of the ACcomponent was varied from 8, 10 and 12% of the

DC level. The magnitude of the dynamic impedanceappears to decrease as the magnitude of the ACcomponent increases. The total deviation over thisrange of tipple amplitudes is on the order of .25 dB.

The results of the phase angle measurements weresimilar.

The effect of the DC current level on the dynamicimpedance was investigated with arc currents of 8

and 10 amperes respectively. The magnitude of theAC component was maintained at 10% for the

investigation. The static impedance data in Figure 3show that the slope of the static volt-ampere curvewas more negative at 8 than at 10A. Thus, it wasexpected that the magnitude of the dynamicimpedance at 8A would be greater than at the 10Apoint. This indicates that the gain constant C ofequation 9 is also dependent on the DC current level.The impedance magnitude data are presented inFigure 8a. At 8 amperes the magnitude of thedynamic impedance was in fact an average of 4 dBgreater than at the 10 ampere level. The impedancephase data are presented in Figure 8b. The phaseangle for the 8A case was virtually identical to the10A case.

Thruster Performance

The effects of power processor ripple on arc jetperformance had not been determined. To quantifythese effects, the thrust produced by the arc jet wasmeasured while varying the tipple frequency. Forthese measurements, the ripple peak-to-peak

amplitude was maintained at 10% of the I)(2 value.

One possible scenario for a degradation inperformance was a change in the dissipated power in

the arcjet with tipple present when compared to theDC case. This was supported by previous workwhich demonstrated that the pressure in the arcchamber of the thruster was proportional to the

input power, and therefore, thrust and specificimpulse were also related to the input

power24.When calculating the power dissipated inthe DC case, only the static impedancecharacteristics were of interest, thus the powerdissipated was simply the product of the DC arcvoltage and current. However, when the AC tipplecomponent was superimposed, the dynamicimpedance was also considered. If the current inputto the arcjet was:

iarc= iac + iacsin (2rift) (i0)

where iac was the amplitude of the ripple, or one

half of the peak-to-peak amplitude, then theresulting arc voltage was:

Var c = Vdc+ vacsin (2rift + 0) (11)

where 0 was the phase angle between the AC arcvoltage current, using the current as the reference.The amplitude of the AC arc voltage and the phaseangle 0 were determined from the dynamicimpedance data presented in Figures 4a and b. The

dynamic impedance magnitude data of Figure. 4awere measured by the gain/phase meter as the ratioof the amplitudes of the AC arc voltage over the ACarc current, represented by equation 3. Since theamplitude of the AC arc current was 1A peak-to-peak, the AC voltage peak-to-peak amplitude wasexactly equal to the dynamic impedance in ohms,

which could be calculated using equation 8. Thispeak-to-peak amplitude was then divided by 2 todetermine the actual amplitude of the AC arc

voltage. The phase angle 0 was taken directly fromthe data of Figure 4b, and the RMS powerdissipation was then calculated with the relation:

Vaclac ._x

Pnns = Vdcidc+ _ COS_O) (12)

Since the dynamic impedance had a magnitudeand phase, a real and complex or imaginarycomponent were associated with the impedance. Ifthe value of the real component of the impedance

was negative, the value of cos0 was also negative.This was the case for the entire frequency range of

this experiment since 900<0<270 ° as shown inFigure 4b. The tight hand term of equation 12 thenconstituted a loss when compared to the pure DCcase, since the DC current level and DC arc voltageremained constant. The magnitude of the lossdepended on the magnitude of the AC V-I productand the value of the phase angle.

The experiment apparatus allowed thesimultaneous measurement of the thrust, AC and

DC are voltages and currents, and the flow rate.Thus, the thrust and power dissipated could becontinuously monitored while the tipple frequency

wasvaried.Theimpedancedatatakenwhileon thethrust stand correlated with the impedance datashown in Figures 4a and 4b, taken in a separatefacility. In the frequency range tested, the value for 0

remained greater than 90 ° , thus the value of theright hand term of equation 12 remained negativethroughout. However, the magnitude of the V*I

product for the AC component was small, in thiscase less than 5 watts, which is less than 1% of the

total power.The thruster was operated at a fixed flow rate with

a DC current level of 10A with an AC component

of 1Ap_p. Figure 9 presents the thrust data for threeflow rates. The DC thrust levels for each flow rate

are represented by a horizontal line. The thrust levelremained constant throughout the frequency rangetested, therefore, the specific impulse also remainedconstant. The overall efficiency is presented in

Figure 10. There are minor deviations in theefficiency, but these are more likely due to thepropagation of measurement errors in thecalculations. The changes in performance wereinsignificant at all frequencies tested, as wasexpected from the power calculations.

CONCLUSIONS

The interface between a 1 kW class arc jet and

future power processors was characterized bydetermining the impacts of power supply outputcharacteristics on arcjet performance and by

characterizing the arcjet as a static and dynamicelectrical load. The impacts of power supply rippleon arcjet performance were assessed with a ripple

amplitude of 10% peak-to-peak of the DC currentlevel.

Characterization of the arc jet as an electrical loadrequired the measurement of the static and dynamicimpedances of the arc. The static impedance datashow that the DC operating voltage of the arc jet is afunction of the input current and flow rate. The arcvoltage decreases with increasing current andincreases with flow rate. A curve fit of the V/I data

results in an equation that fairly accurately describesthe static impedance for a given flow rate, providedthat the cathode has been run for a sufficient time torender the effect of further erosion negligible.Variations in the static impedance from thruster tothruster were not characterized.

The dynamic impedance also exhibits similartrends with flow rate. Initially it was intended toassign poles and zeros to the impedance usingfrequency response methods. However, from thesedata it was difficult to assign poles and zeros to thedynamic impedance because the magnitude remainedconstant to within 2 dB across the 100 Hz to 100

kHz range tested. A phase shift of approximately

180 ° was noted at low frequencies, but decreased at

frequencies above 10 kHz. This is representative of atransportation lag in the arc column on the order of1.5 _. The transportation lag was multiplied by a

negative gain constant to account for the 180 ° phaseshift at low frequencies. The value of the gainconstant was found to be dependent on flow rate andthe DC current level. Thus, for the frequency rangein tested the dynamic impedance of the arc can be

represented as.(15x16_)i

zd(s) =-C e

A typical value for C was found to be on the orderof 3.76 for a 10A DC current level. This valueincreased as the DC current level decreased for a

given ripple amplitude of 10% peak-to-peak.This magnitude, however, is greater than that

expected from the static curves. This may indicatepresence of a pole-zero pair at a very low frequency.Further work is necessary to investigate this

phenomenon. From these data, there was noindication that the step response of the power

processor regulator should be changed. These datawere taken with two thrusters of nearly identical

configuration and the data compared favorably,indicating that these data were transportable betweenthrusters.

The effects of ripple on arc jet performance wereassessed by measuring DC performance levels atvarious flow rates to provide a baseline forcomparison to thruster performance with powersupply ripple present. An AC component ofconstant amplitude and variable frequency was theninjected to simulate power supply ripple. Thrust andinput power were measured simultaneously whilethe ripple frequency was varied from 100 Hz to 100kHz. The dynamic impedance of the arc jet caused thedissipated power in the arcjet to decrease slightly dueto the phase angle associated with the dynamic

impedance which was greater than 90 ° . The totalreduction in power was less than 1%, thus, the AC

component was found to have no measurable effecton thrust, specific impulse or overall efficiency overthe frequency range tested. This indicates that presentpower supply designs, with output ripple currentson the order of 10% peak-to-peak, are notcompromising overall system performance. Theimpact of ripple on thruster lifetime was notassessed in this work, however, past experienceindicates that this is not an issue for mission

lifetimes of 1000 hours.

REFERENCES

1. John, R.R., Connors, J.F., and Bennet, S.,"Thirty Day Endurance Test of a 30 kW ArcjetEngine", AIAA Paper 63-274, June 1963.

7

2. John, R.R., "Thirty Kilowatt Plasmajet Rocket-Engine Development," RAD-TR-64-6, Avco Corp.,Wilmington, MA, NASA CR-54044, 1964

3. Todd J.P., and Sheets, R.E., "Development of aRegeneratively Cooled 30-kW Arcjet Engine,"AIAA Journal, Vol. 3, No. 1, Jan. 1965, pp. 122-126.

4. Todd, J.P., "30 kW Arcjet Thruster Research,"APL-TDR-64-58, Gianni Scientific Corp., SantaAna, CA, Mar. 1964. (Avail. NTIS, AD-601534.)

5. Ducati, A.C., Humpal, H., Metzler, J.,Muehlberger, E., Todd, J.P., and Waltzer, H., "l-kW Arcjet-Engine System-Performance Test,"Journal of Spacecraft and Rockets, Vol. 1, No. 3,

May-June 1964, pp. 327-332.

6. Mc Caughey, O.J., Geideman, W.A., Jr., amdMueller, K., "Research and Development of a 2 kWArc-Jet Thruster," GRC- 1646, Plasmadyne Corp.,Santa Ana, CA, NASA CR-54035, 1963.

7. Shepard, C.E., and Watson, V.R., "Performanceof a Constricted-Arc Discharge in a Supersonic

Nozzle," Physico-Chemical Diagnostics of Plasmas,T.P. Anderson, R.W. Springer, and R.C. Warder,Jr., eds., Northwestern University Press, 1964, pp.261-272. (AIAA Paper 63-380).

8. Wallner, L.E., and Csika, J. Jr., "Arc-Jet

Thrustor for Space Propulsion," NASA TN D-2868,1965.

9. Nakanishi, S., "Experimental performance of a 1Kilowatt Arcjet Thruster," AIAA Paper 85-2033,Oct. 1985. (NASA TM-87131.)

10. Curran, F.M., and Nakanishi, S., ',Low Power

dc Arcject Operation with Hydrogen/NitrogenPropellant Mixtures," AIAA Paper 86-1505, June1986. (NASA TM-87279)

11. Hardy, T.L., and Curran, F.M., "Low Power dcArcjet Operation with Hydrogen/Nitrogen/Ammonia Mixtures," AIAA Paper 87-1948, June1987. (NASA TM-89876)

12. Knowles, S.C., Smith, W.W., Curran, F.M.,and Haag, T.W., "Performance Characterization of a

Low Power Hydrazine Arcjet," AIAA Paper 87-1057, May 1987.

13. Knowles, S.C., "Arcjet Thruster Research andTechnology, Phase I, Final Report," 87-R-1175,

Rocket Research Co., Redmond, WA, Sept. 1987.

14. Simon, M.A., Knowles, S.C., Curran F.M.,and Hardy, T.L., "Low Power Arcjet Life Issues,"AIAA Paper 87-1059, May 1987.

15. Gruber, R.P., "Power Electronics for a l-kW

Arcjet Thruster," AIAA Paper 86-1507, June 1986.(NASA TM-87340)

16. Curran, F.M., and Haag T.W., "ArcjetComponent Conditions Through a Multistart Test,"AIAA Paper 87-1060, May 1987. (NASA TM-89857)

17. Haag, T.W., and Curran, F.M., "Arcjet StartingReliability: A Multistart Test on Hydrogen/Nitrogen Mixtures," AIAA Paper 87-1061, May1987. (NASA TM-89867)

18. Sarmiento, C.J., and Gruber, R.P., "Low Power

Arc jet Thruster Pulse Ignition," AIAA Paper 87-1951, July 1987. (NASA TM-100123)

19. Knowles, S.C.,Arc jet Thruster Research andTechnology Phase II., NASA CR 182276, 1990,(To be published)

20. Curran, F.M., and Haag, T.W., "An ExtendedLife and Performance Test of a Low Power Arcjet,"AIAA Paper 88-3106, July 1988. (NASA TM-100942)

21. Gupta, S.C., and Hasdorff, L., Fundamentals of

Automatic Control, Robert E. Krieger PublishingCo., Florida, 1983

22. Cobine, J.D., Gaseous Cond_l_tors, Dover

Publications Inc., New York, 1941

23. Ogata, K., Modem Control Engineering.

Prentice-Hall Inc., Englewood Cliffs, New Jersey,1970

24. Curran, F.M., "An Experimental Study ofEnergy Loss Mechanisms and EfficiencyConsiderations in the Low Power dc Arcjet," AIAA

Paper 85-2017, September 1985. (NASA TM-87123)

8

INJECTION

FRONT DISK "_

REAR INSULATOR _

INLET CATHODE

Figure 1. - Cutaway view of arcjet thruster.

POWERSUPPLY

150V

ISA

DIODE

CURRENT ___REGULATOR MODULATOR

/I FUNCTIONGENERATOR

CURRENTCOAXI AL

SHUNT PROBECABLE RESISTOR

VACUUM

RELAY I PROBEI PULSE I AMPLIFIERP

[TRANSFORMERIREF, l

I GAIN/PHASE 1"

T?OSCILLOSCOPE/II

COUNTER

VACUUM

FACILITY

ARC JET

I ISOLATIONAMPLIFIER

Figure 2a. - Apparatus block diagram.

9

POWER

SUPPLY

ISOV

1SA

ireg iarc

CURRENT I +REGUL

-- _ imod

FUNCTION I

GENERATOR S VACUUMRELAY

Figure 2b. - Regulator/Modulator operation.

o"

t_

o

125

120

115

1108.0 910 10.0 il.0

Arc Current,A

Figure 3. - Static Volt/Ampere characteristics for various flow rates, Thruster 1.

Mass Flow kg/sec

O 4.8x10 o5

[] 4.5x10 -5

A 4.3x10 -5

10

mqDe"'10

t.-

o=31

12

11

10

[] 0 [] 0 00 0 O0

A A A AAA

• • . • • w,.| • • • =.i|

10 2 10 3 10 4

Frequency, Hz

A

A

[]

&

| • • w •=

0 5

Mass Flow [kg/sec]

O 4.8x10-5

[] 4.5x10 -5

A 4.3x10-5

Figure 4a. - Dynamic impedance magnitude vs. frequency at various flow rates, thruster 1.

es

_et'=g

e-ts

20O

180

160

140

120

100

B B []

[][]

B

[]

= • • • . _.=| . • . . = .m,| = • • • • i i

0 2 10 3 10 4 10 5

Frequency,Hz

Mass Flow,kg/secO 4.8x10 -5

[]4.5x10 -5

A 4.3X10 -5

Figure 4b. - Dynamic impedance phase angle vs. frequency at various flow rates, thruster 1.

11

g0"o@

"o

¢-

14

13

12

11

10

9

0 0 0

0 0 O0 0

[] [] n[] [] [] n n

0

r-l

0

[]

OoC

00

Dr

i , • i , ,,w| , • , roll| w • w , , , ,

10 2 10 3 10 4 10 5

Frequency, Hz

0 Thruster2

[] Thruster 1

Figure 5a. - Comparison of dynamic impedance magnitude, thrusters 1 and 2, flow rate and DC current fixed.

220

2OO

180

@160

,<

i/)

" 1400,.

120

100

0019 8 813 [] []

P'B

• i , , , , m| i , , ,l| |

10 3 10 4

Frequeney, i_Z

0

[]

B0

%

, i i i I,

O 5

0 Thruster2

[] Thruster 1

Figure 5b. - Comparison of dynamic impedance phase angle, thrusters 1 and 2, flow rate and DC currrent fixed.

Ill"OQ-

t_

12.0

11.5

11.0

10.5'

10.0'

9.5'

9,0'

8.5'

8.0

FI FI I-I r"l I"1 - I

[((((l[ [[((((( tt[([[ [[(I.t(|/[(((( tin I[([ (([f[(([¢[£[C( ((] I[(( [[[([[ [[(({[[ [[(([[[ [[[[[ i

/121

= i • , • ,, ,| , • , | i i, w| • • i m i

0 2 10 3 10 4 10 5

Frequency,Hz

O Model

[] Thruster

Figure 6a. - Comparison of dynamic impedance magnitude predicted by model and thruster data, .00048 kg/sec flow rate.

e-<

t_e-D.

2OO

150

100

50

0 0 0 0 0 0

I IIIilli IIIIhq IIIIII III1111 II II_'l _lllh I II1_

• ' J''l " ' ' "vTI ' ' ' ' . .'

0 2 10 3 10 4 10 5

Frequency, Hz

0 Thruster

13 Model

Figure 6b. - Comparison of impedance phase angle predicted by model and thruster data, .00tN8 kg/sec flow rate.

13

m0""O

t_

12

11

10

- 00 0

0 A0 [] n

0 0 A []

[]

0

[] 0

[]A

0

[]A

0

A

[]

' • ° ' °'I ' ° ° °'|I ° " ' ''

0 2 10 3 10 4 10 5

Frequency, Hz

Ripple Amplitude

O 8o/0

[] 10%

A 12%

Figure 7a. - Dynamic impedance magnitude vs. frequency for various peak-to-peak ripple amplitudes, thruster 2.

6_Q

O')

.<

r-at

2OO

180

160

140

120

100

0 2 10 3 10 4 10 5

Frequency,Hz

Ripple Amplitude

0 80/0

[] 10%

A 12%

Figure 7b. - Dynamic impedance phase angle vs. frequency for various ripple amplitudes, thruster 2.

14

"06

==t,-

18

16

14

12

10

8

0 0 O0

00 0

0

[] [] on [] [] DO

0

[]

0

[]

0

[]

• , , ,,,| , , , , J, | , , , | ,r

0 2 10 3 10 4 10 5

Frequency, Hz

DC Current

0 8A

[] 10A

Figure 8a. - Dynamic impedance magnitude vs. frequency for different DC current levels, thruster 2.

O

,<

t_

f

2O0

180

160

140

120

[]

fl [] Oo0

0

8fl

100 ..... , ...... ,

10 2 10 3 10 4

fl

8

O

[3

DC Current

O 8A

[] 10A

0 5

Frequency, Hz

Figure 8b. - Dynamic impedance phase angle vs. frequency for different DC current levels, thruster 2.

15

ZE

t==J_p-

22O

I215 _ 0 0 04 )

210 t

[] [] 13 1

205

200

195

190 ..... , ...... ,10 2 10 3 10 4 10 5

Mass Flow, kg/sec

O 4.8x10 -5

[] 4.5x10 -5

A 4.3x10 -5

Frequency, Hz

Figure 9. - Thrust vs. ripple frequency for 10% peak-to-peak tipple amplitude. Horizontal lines denote DC performance levels.

0.41

0.40'

oc =e

I.U

0.39 -

0.38

O 2

[] [] []0

A

• , ,,,| ..... ,!

10 3 10 4

Frequency, Hz

A A

• , , • , , ,,

O5

Mass Flow, kg/sec

O 4.8x10 -5

[] 4.5x10 -5

A 4.3x10 -5

Figure 10. - Efficiency vs. tipple frequency for 10% peak-to-peak tipple amplitude, thruster 1.

16

Report Documentation PageNational Aeronautics andSpace Administration

1. Report No. NASA TM-103190 2. Government Accession No, 3. Recipient's Catalog No.

AIAA-90-2579

5. Report Date4. Title and Subtitle

Arcjet Load Characteristics

7. Author(s)

John A. Hamley

9, Performing Organization Name and Address

National Aeronautics and Space AdministrationLewis Research Center

Cleveland, Ohio 44135-3191

12. Sponsoring Agency Name and Address

National Aeronautics and Space Administration

Washington, D.C. 20546-0001

6. Performing Organization Code

8. Performing Organization Report No.

E-5578

10. Work Unit No,

506-42-31

11. Contract or Grant No.

13. Type of Report and Period Covered

Technical Memorandum

14. Sponsoring Agency Code

15 Supplementary Notes

Prepared for the 21st International Electric Propulsion Conference

JSASS, Orlando, Florida, July 18-20, 1990.

cosponsored by the AIAA, DGLR, and

16 Abstract

Experiments were conducted to define the interface characteristics and constraints of 1 kW class arcjets run on

simulated decomposition products of hydrazine and power 9rocessors. The impacts of power supply output

current ripple on arcjet performance were assessed by variation of thc ripple frequency from 100 Hz to 100 kHz

with 10 % peak-to-peak ripple amplitude at 1.2 kW. Ripple had no significant effects on thrust, specific impulse

or efficiency. The impact of output ripple on thruster lifetime was not assessed. The static and dynamic

impedances of the arcjet were quantified with two thrusters of nearly identical configuration. Superposition of an

AC cornponent on the DC arc current was used to characterize the dynamic impedance as a function of flow rate

and DC current level. A mathematical model was formulated from these data. Both the static and dynamic

impedance magnitude were found to be dependent on mass flow rate. The amplitude of the AC component was

found to have little effect on the dynamic impedance. Reducing the DC level from 10 to 8 amps led to a large

change in the magnitude of the dynamic impedance with no observable phase change. The impedance data

compared favorably between the two thrusters.

17. Key Words (Suggested by Author(s))

Space propulsion

Electric propulsion

Arciet thrusters

Power processing

18. Distribution Statement

Unclassified - Unlimited

Subject Category 20

19. Security Classif. (of this report) 20. Security Classif. (of this page) 21. No, of pages

Unclassified Unclassified 18

NASAFORM_62eOCT86 *For sale by the National Technical Information Service, Springfield, Virginia 22161

22. Price"

A03

7 7 _