University of Southern California Dissertations and Theses

Development of the two-stage micro pulsed plasma thruster

(USC Thesis Other)

Development of the two-stage micro pulsed plasma thruster

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

DEVELOPMENT OF THE TWO-STAGE
MICRO PULSED PLASMA THRUSTER

by

John Schilling

________________________________________________________

A Dissertation Presented to the
FACULTY OF THE VITERBI SCHOOL OF ENGINEERING
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(ASTRONAUTICAL ENGINEERING)

May 2007

Copyright 2007 John Schilling
ii
Acknowledgements

I would like to offer my most sincere gratitude to Dr. Dan Erwin for his encouragement,
guidance, and especially his infinite patience, without which none of this would be possible.
The same goes to Drs. Muntz, Kuncs, Gundersen, and Katsoulas, and the rest of the
excellent faculty and staff at USC. Thanks for sticking with me all these years. I’m not sure
if I’ve set a department record here, but it must be close. Does anyone even remember when
my dissertation topic was pulsed electron beam fluorescence?

I would also like to thank Dave White, Stu Bushman, and Garrett Reed of W.E. Research,
LLC. Good friends all, and good partners during the difficult years of getting this business
off the ground. None of this would have happened without Dave’s initiative and willingness
to buck the system. That Stu was as talented a business manager as he was a propulsion
engineer, surprised us all and on more than one occasion saved us all. And anyone who
needs an automated diagnostic suite in their lab, or a companion for an expedition to Alaska
or Peru, would be hard-pressed to find anyone better than Garrett.

Thanks also to the Air Force Research Laboratory and the Air Force Office of Scientific
Research for sponsoring this work. Particularly Ron Spores and Greg Spanjers for their
strong support in the earliest stages of this enterprise. They’ve both moved on, but their
legacy remains. And credit goes to Captain James Lake, USAF, and Dr. Mike Dulligan of
AFRL, for support in the laboratory. Thanks to them, micronewton thrust measurements are
a push of a button, not tedious weeks of effort. Mike was also my copilot on many flights
over Southern California, it is both appropriate and tragic that he passed away in a light
plane crash a few years ago.

Finally, kudos to the United States Air Force Academy, particularly Major Tim Lawrence
and Lt. Pam Fetchko, and to Vlad Hruby and Bruce Pote at Busek. Building a thruster is an
excuse to play in the lab and feel important. Building a spaceship, is the real deal, and flying
the μPPT on FalconSat 3 is what makes this worthwhile.

iii
Table of Contents

Acknowledgements. . . . . . . ii

List of Tables . . . . . . . v

List of Figures . . . . . . . vi

Abbreviations . . . . . . . x

Abstract . . . . . . . xi

Chapter I: Introduction . . . . . 1

Chapter II: The Pulsed Plasma Thruster . . . 6
2.1 PPT Operation . . . . . 7
2.2 Typical Parameters . . . . . 9
2.3 The Triggered μPPT . . . . 12
2.4 Performance . . . . . 14

Chapter III: Mission Applications . . . . 17
3.1 TechSat 21 Spacecraft Design . . . 19
3.2 Micropropulsion Options . . . . 25
A: Chemical Micropropulsion . . . 27
B: Electromagnetic Micropropulsion . . 29
C: Electrostatic Micropropulsion . . 30
D: Electric Power Processing . . . 31
3.3 Analysis . . . . . . 33
3.4 Conclusions . . . . . 40

Chapter IV: PPT Performance Improvement . . . 43
4.1 Pulsed Operation – Loss Mechanisms . . 43
4.2 Pulsed Operation – Possible Solutions . . 44
4.3 Propellant Utilization Inefficiencies . . . 47
4.4 Propellant Inefficiency – Possible Solutions . . 50
4.5 The XPPT-48 Testbed . . . . 52
4.6 Alternate Propellant Operation . . . 55
4.7 Effect of Variable Capacitance and Discharge Energy 58
4.8 Conclusions . . . . . 61

Chapter V: The Two-Stage μPPT . . . . 64
5.1 Two-Stage μPPT Concept . . . . 65
5.2 Two-Stage μPPT Fabrication. . . . 68
5.3 Test Facilities . . . . . 72
5.4 The MicroNewton Thrust Stand . . . 76
iv

Chapter VI: Two-Stage μPPT Performance . . . 83
6.1 Thruster Life . . . . . 83
6.2 Propellant Consumption . . . . 86
6.3 Discharge Energy Variation . . . . 88
6.4 Thrust Measurements . . . . 93
6.5 Performance Summary . . . . 97

Chapter VII: μPPT Plume Studies . . . . 99
7.1 Near-Field Plume Diagnostics . . . 100
7.2 Herriot Cell Interferometer . . . . 101
7.3 Near-Field Density Measurements . . . 103
7.4 Thruster and Model Comparison . . . 104
7.5 Plume Contamination Study . . . . 106
7.6 Plume Contamination Results . . . 108

Chapter VIII: Advanced μPPT Concepts . . 113
8.1 Chemically Energetic Propellants . . . 113
8.2 Thinned Outer Electrodes . . . . 114
8.3 Clustered μPPT . . . . . 117
8.4 Multi-Axis Switching of the μPPT . . . 120

Chapter IX: The μPPT Flight Unit . . . . 126
9.1 TechSat 21 First Redesign . . . . 126
9.2 New Performance Requirements . . . 130
9.3 TechSat 21 Second Redesign . . . 133
9.4 FalconSat 3 . . . . . . 137
9.5 Delivery of the Flight Unit . . . . 139

Chapter X: Conclusions . .. . . . 141

Bibliography . . . . . . . 144

v
List of Tables

Table Page
2.1 Thruster Performance Characterization . . . 15
3.1 Micropropulsion Options . . . . 32
3.2 TechSat 21 Sample Propulsion System Mass Breakdown . 36
3.3 TechSat 21 Propulsion System Mass (Flight Experiment) . 37
3.4 TechSat 21 Propulsion System Mass (Operational Mission) 38
3.5 TechSat 21 Propulsion Options Comparison . . 40
4.1 XPPT-48 Performance . . . . . 55
4.2 Ablation Rates as a Function of Power Level . . 59
5.1 Baseline Two-Stage μPPT Configuration . . 68
5.2 Thrust Stand Performance Study . . . . 82
6.1 μPPT Overall Performance . . . . 97
9.1 Mass Breakdown for TechSat 21 First Redesign μPPT System 130

vi
List of Figures

Table Page
2.1 Conventional PPT Schematic . . . . 6
2.2 The LES 8/9 PPT . . . . . 10
2.3 The LES 8/9 Trigger Circuit . . . . 11
2.4 The Triggered μPPT . . . . . 12
2.5 Triggered μPPT Schematic . . . . 13
3.1 TechSat 21 Spacecraft . . . . . . 20
3.2 TechSat 21 Propulsion System Layout . . . 36
4.1 Rectilinear vs. Coaxial Magnetic Field Geometry. . 45
4.2 Alternate PPT Circuit Schematics . . . 46
4.3 Alternate PPT Circuit Behavior . . . . 46
4.4 XPPT-48 Cross Section . . . . . 53
4.5 Assembled XPPT-48 Thruster . . . . 54
4.6 XPPT-48 Thruster Schematic . . . . 54
4.7 XPPT-48 Current Traces, 19.7 μF . . . 60
4.8 XPPT-48 Current Traces, 42.0 μF . . . 60
4.9 Higher Energy XPPT-48 Current Traces . . . 61
4.10 PPT Performance Trends . . . . 63
5.1 Two-Stage μPPT . . . . . 65
5.2 Two-Stage μPPT Schematic . . . . 66
5.3 Energizing the Two-Stage μPPT Capacitors . . 67
5.4 μPPT Tube Custom-Made by Precision Tube Company . 69
vii
5.5 Custom Electronics, Inc Capacitors . . . 70
5.6 EMCO E40 4.0 kV Power Supply . . . 71
5.7 Baseline Two-Stage μPPT . . . . 72
5.8 Chamber 8 (Glass Bell Jar) . . . 73
5.9 Chamber 9 (Steel Bell Jar) . . . . 74
5.10 Thermal Shroud in AFRL Chamber 5A . . . 75
5.11 Chamber 2 with Thrust Stand . . . . 76
5.12 AFRL Forced-Harmonic Oscillator Thrust Stand . . 77
5.13 Thrust Stand Oscillation, Amplitude, and Long-Term Drift 78
5.14 Calibration Behavior . . . . . 79
5.15 Resulting Linear Calibration . . . . 80
6.1 Two-Stage μPPT Fires Until Propellant Depletion . 84
6.2 Char Formation in a Two-Stage μPPT . . . 85
6.3 Propellant Consumption Over Long-Duration Firings . 86
6.4 Pockels Cell Schematic . . . . . 89
6.5 Pockels Cell Setup . . . . . 90
6.6 2.25J, 2 Hz μPPT Breakdown Voltage . . . 91
6.7 Breakdown of a μPPT Fired for 350,000 Seconds at 6 Joules . 92
6.8 Pulse Energy of a μPPT Fired for 350,000 Seconds at 6 Joules 92
6.9 Modified Baseline μPPT Performance (2.25 J) . . 94
6.10 Low-Impedance μPPT Performance (1.96 J) . . 95
6.11 Low-Impedance μPPT Performance (1.00 J) . . 95
6.12 Low-Impedance μPPT Performance (0.64 J) . . 96
7.1 Interferometer Setup . . . . . 102
viii
7.2 Electron Density for 13-Pass Herriot Cell . . 103
7.3 Neutral Density for 13-Pass Herriot Cell . . . 104
7.4 Predicted vs. Measured Electron Density . . . 105
7.5 BOL Plume, No Recession . . . . 108
7.6 BOL LTV, No Recession . . . . 109
7.7 7mm Recession Plasma . . . . . 109
7.8 7mm Recession LTV . . . . . 110
7.9 Geometric LTV Divergence Model . . . 110
7.10 17.5mm Recession Plasma . . . . 111
7.11 17.5mm Recession LTV . . . . . 111
7.12 54mm Recession Plasma . . . . 111
7.13 54mm Recession LTV . . . . . 112
7.14 Divergence of Plasma and LTV Plumes . . . 112
8.1 Standard vs. Thin-Walled μPPT . . . . 116
8.2 Cluster of Two-Stage μPPT Tubes . . . 117
8.3 Recessed Propellant in a Cluster of Two-Stage μPPT Tubes 118
8.4 Multiple Thrust Axis Switching . . . . 121
8.5 Optically Isolated High-Voltage Switch . . . 123
8.6 Switched μPPT Performance . . . . 124
9.1 TechSat 21 First Redesign . . . . 127
9.2 μPPT First Redesign . . . . . 128
9.3 μPPT Functional Diagram . . . . 129
9.4 TechSat 21 Second Redesign . . . . 134
9.5 μPPT Demonstration Unit . . . . 135
ix
9.6 μPPT Demonstration Unit Block Diagram . . 137
9.7 Flight-Like μPPT on the USAFA Vibration Table . 138
9.8 Busek Prototype μPPT Flight Unit . . . 140

x
Abbreviations

μPPT Micro Pulsed Plasma Thruster
ACS Attitude Control System
AFRL Air Force Research Laboratory
BOL Beginning of Life
BHT Busek Hall Thruster
DSMC Direct Simulation Monte Carlo
EOL End of Life
FEEP Field Effect Electrostatic Propulsion
HAN Hydroxyl Ammonium Nitrate
HET Hall Effect Thruster
LES Lincoln Experimental Satellite
LTV Late Time Vaporization (and particulate emission)
MEMS Micro ElectroMechanical System
PIC Particle in Cell
PPT Pulsed Plasma Thruster
PPU Power Processing Unit
USAF United States Air Force
USAFA United States Air Force Academy
WER White Engineering & Research, LLC
XPPT Experimental Pulsed Plasma Thruster

xi
Abstract

In this dissertation, the development of the two-stage micro pulsed plasma thruster (μPPT)
is described. This development followed from a research effort aimed at overcoming
crippling performance limitations of the conventional PPT. Several fundamental causes of
this performance shortfall were identified, and attempts to overcome them were uniformly
unsuccessful. Furthermore, a trade study of other micropropulsion systems suggested that
even an improved conventional PPT would find little use.

However, previous work had suggested that the conventional PPT could be greatly reduced
in size and complexity, and the same mission-specific trade study indicated that a micro PPT
would be quite useful even if of limited performance. Previous testbed μPPTs, operating in
an actively triggered mode, were laboratory curiosities of little practical value. The two-
stage μPPT, with triggering performed in a separate stage, passively coupled to the main
discharge, greatly simplifies the device while allowing a much larger propellant loadout and
total propulsive impulse. A complete three-axis thruster module with 25 N-s of impulse
along each axis, can be constructed from off-the-shelf components with a total mass of less
than one kilogram.

The two-stage μPPT was developed and extensively tested, proving to be a robust, reliable
system with sufficient performance for near-term missions. Several avenues for further
improvement were identified. A flight μPPT has been constructed, representing the current
state of the art in this area, and will fly on the USAF FalconSat-3 spacecraft in early 2007.
1

Chapter I

Introduction

While rocket propulsion remains unsurpassed for the efficient generation of high thrust for
spacecraft propulsion, it is fundamentally limited by the energy densities of chemical fuels –
no more than 15 MJ/kg even for the most energetic propellants. Higher energy densities can
be achieved by using an external source of energy to accelerate an inert propellant. Since
most spacecraft can generate nearly unlimited amounts of electric energy by means of solar
arrays and/or nuclear power sources, there is the potential for a substantial increase in
propulsive efficiency.

The peak power limits of solar or even nuclear electric power sources limit such electric
propulsion systems to relatively low thrust. Electric spacecraft propulsion will likely never
be suitable for space launch applications. However, most on-orbit maneuvers do not face
any intrinsic time limit. Where such maneuvers can be conducted over an extended period
of time, weeks or months, the use of electric propulsion systems can result in substantial
reductions in propellant consumption. This translates directly to an increase in spacecraft
payload and/or endurance, and thus to increased mission capability.

For the first thirty or so years of the space age, the general tendency was for spacecraft to
increase in size, power, and capability. The history of electric propulsion over that period
mirrored this trend, culminating with the operational use of ion and Hall effect thrusters at
power levels of up to 4.5 kW, with even more powerful systems still in development.
However, over the last ten years, there has been a renewed interest in very small spacecraft,
2
of under 100 kg total mass. These so-called microsatellites are individually of limited
capability, but can be deployed in large numbers. For some applications, a large number of
cooperative microsatellites seems to offer superior performance to the traditional solution of
one large spacecraft.

This renewed interest in small spacecraft, calls for the development or redevelopment of
small, low-power electric propulsion systems. One candidate technology is the Pulsed
Plasma Thruster, or PPT. The PPT was originally developed in the late 1960s, but offered
only mediocre performance and could not be efficiently scaled to high power levels. It thus
fell out of favor fairly quickly with the general increase in spacecraft mass and power, but
has recently been subject to renewed interest.

Chapter II of this thesis offers an overview of PPT development, addressing the history,
design, operating principles, and performance of the device. It also introduces the micro
PPT, or μPPT, a recent development that greatly reduces the mass and complexity of the
PPT.

The PPT and μPPT are only two among many promising micropropulsion technologies. In
order to prioritize research and development efforts in this area, a detailed trade study of
micropropulsion systems at various levels of technical maturity was conducted by this
author, for the specific purpose of selecting a propulsion system for the USAF’s TechSat 21
microsatellite. This was a fairly ambitious program involving a number of small spacecraft
operating in close formation, with propulsive requirements beyond the capability of existing
propulsion systems.
3

Chapter III of this thesis describes this trade study in some detail, and supports two
conclusions. First, that the conventional PPT is not a promising technology unless the
performance can be substantially increased. Second, that the μPPT may represent an
absolute minimum-mass and minimum-complexity solution for providing propulsive
capability to a spacecraft, and thus be an enabling technology for very small spacecraft.
However, previous μPPT designs were limited to extremely low propellant loads and total
propulsive impulse, making them suitable only for the least demanding missions.

As a result of this study, substantial attention was devoted to PPT and μPPT technology
development by the Air Force Research Laboratory. The W.E. Research Corporation, of
which this author is chief scientist, was contracted to perform a four-year effort to advance
PPT and μPPT technology with the aim of producing propulsion systems for the TechSat 21
mission. This effort represented the bulk of WER’s activities between 1999 and 2003.

An exploration of the causes, and potential remedies, for the low performance of the
conventional PPT is given in chapter IV. This represented a substantial research effort and
included the development, construction, and testing of a new PPT design. Several different
thruster geometries, operating modes, and propellants, were tested. Unfortunately, this effort
proved fruitless. Rather than improving the performance of the PPT, the result was to
solidify the conclusion that the limitations of the PPT are rooted in fundamental physical
limits that seem likely to defy technological solution for the foreseeable future, and that
future effort at PPT development is likely to represent wasted effort.

4
The μPPT, however, does remain a fruitful area for research. Its performance is limited by
the same physical effects as that of the conventional PPT, but its low mass and complexity
allow it to be used in applications where no other propulsion system is available. If the
choice is between a mediocre propulsion system and none whatsoever, the mediocre
propulsion system is a clear winner. It is therefore worth the effort to develop the best
possible μPPT system

Chapter V, the heart of this thesis, details the development and construction of the two-stage
μPPT, a device that overcomes some of the unique limitations of the original μPPT design
and allows performance comparable to a conventional full-scale PPT to be achieved in the
much smaller package of an μPPT. The functionality and performance of this new thruster
are subject to extensive testing, as described in Chapter VI.

Chapter VII is a digression into the area of plume diagnostics and modeling. In one respect,
the plume of a thruster is of supreme unimportance – thruster exhaust is generally supersonic
at the exit plane regardless of the propulsion technology, disturbances cannot propagate
upstream, and so nothing that happens in the plume can affect the operation or performance
of the thruster. However, direct observation of events and conditions interior to the thruster,
which can affect its operation, is frequently difficult, and the plume may be studied for what
insight it may offer to the upstream environment in which it was formed. Of less scientific
but more practical importance, the nature and spatial extent of the plume is of great
importance to spacecraft designers, who must ensure that the plume does not erode,
contaminate, or otherwise damage any important portion of the spacecraft.

5
Chapter VIII is a discussion of three possible techniques for further improving the capability
of the μPPT. Preliminary experiments demonstrated that an alternate propellant promising
increased performance are suitable for μPPT operation, that multiple μPPT modules can be
coupled in parallel to a single power supply for increased total impulse at no cost in system
complexity, and that switching between multiple μPPT modules can be accomplished
relatively simply such that a single system can selectively provide thrust along multiple axes.
Full development of these techniques was beyond the scope of this effort.

Finally, Chapter IX describes the development, manufacture, and testing of an actual flight
PPT unit. While changing priorities in the intervening years ultimately resulted in the
cancellation of the TechSat 21 mission, the μPPT was ultimately selected as the sole attitude
control actuator for the FalconSat 3 spacecraft. This mission represents the final proof of the
μPPT concept, an actual space flight. Flight hardware was delivered earlier this year, and
launch is expected early in 2007.

6

Chapter II

The Pulsed Plasma Thruster

The Pulsed Plasma Thruster is a solid–state electromagnetic thruster suited for on–orbit propulsion
of small spacecraft. The basic mechanism of a PPT consists of two electrodes positioned so as to
form a capacitor–driven arc discharge across the face of a solid propellant bar. The discharge
ablates and ionizes a small quantity of propellant from the face, and the discharge current flowing
through this plasma interacts with its self–generated magnetic field to produce a JxB force which
accelerates the plasma towards the thruster exit, producing thrust
7
.

Figure 2.1: Conventional PPT Schematic
CAPACITOR
ANODE
PLASMA
CURRENT
CATHODE
SPARK PLUG
TRIGGER
CIRCUIT
TEFLON
PROPELLANT
BAR
SPRING
7
The pulsed plasma thruster was first developed in the 1960s, and was used operationally as early
as 1964 on the Russian Zond–2 mission. The US first used PPTs on the TIP/NOVA vehicles in
1976. Early designs were plagued by extremely low electrical efficiencies, usually less than ten
percent, but did demonstrate specific impulse values of order 1000 seconds. This high specific
impulse, along with the extreme simplicity of a propulsion system with an inert solid propellant
and no moving parts, recommended the PPT for attitude control and stationkeeping use in small,
low–cost spacecraft
1
.
With the general growth in spacecraft size and complexity during the 1970s and 1980s, the PPT
lost favor, and no great effort was made to overcome the efficiency problems of the early (and still
largely unchanged) designs. However, the current shift towards the use of smallsats and microsats
in the <100 kg range, alone or in formations, suggests future applications for the PPT. Few high–
I
sp
propulsion systems can be efficiently scaled down to the size and power levels required, and
with even modest improvements the PPT might be ideally suited for such applications
14
.
2.1 PPT Operation
The normal operation of a pulsed plasma thruster begins with the charging of the primary and
trigger capacitors. This is performed by a solid-state high voltage power supply driven by the
spacecraft main bus. The full voltage of the primary capacitor, usually 1-2 kV is applied to the
PPT electrodes, and the insulating nature of the Teflon propellant maintains an open circuit.
When a thrust impulse is required, a trigger pulse is applied to the SCR or other semiconductor
switch in the trigger circuit. This discharges the trigger capacitor through a step-up pulse
transformer. The output of the pulse transformer is applied to the spark plug in the PPT cathode,
where the instantaneous voltage is high enough to initiate a vacuum arc at the spark plug face.
8
This vacuum arc provides a seed plasma sufficient to allow a surface discharge across the
propellant surface separating the PPT anode and cathode. Radiative heating from the arc
discharge immediately adjacent to the propellant surface, and in the earliest stages possibly also
electron or ion collisional heating at the surface, provides energy to ablate propellant from the
surface. The ablated propellant vapor is ionized in the discharge, providing additional charge
carriers. This allows increased current flow, increased radiant arc heating, and thus increased
propellant ablation, which cycle continues until a quasi-steady equilibrium is reached where
propellant is ablated at a rate sufficient to provide charge carriers for a maximum-current (as
limited by circuit inductance and resistance) arc discharge.
This high-current pulsed discharge generates a corresponding pulsed magnetic field, locally
perpendicular to the propellant face and to the discharge current. An examination of the system
geometry will show that the JxB force produced by the interaction between the discharge current
and the magnetic field, will accelerate the charge carriers (i.e. the ablated propellant) away from
the propellant face. This is the primary thrust mechanism of the PPT.
The time constant of the discharge circuit is typically on the order of a microsecond, at which
point the discharge capacitor is exhausted. However, the discharge circuit has a significant
inductance, and so current continues to flow through the circuit and the capacitor is recharged to a
negative voltage whose magnitude is nearly that of the initial charged state. Residual propellant
vapor from the initial discharge allows the arc to be restruck in this reverse configuration, and
indeed several cycles of ringing are observed. These are similar to the original discharge, albeit
with reduced voltage, current, propellant ablation, and so thrust impulse. Eventually, the
maximum voltage in the circuit drops to a level insufficient to strike or maintain an arc, and
current flow ceases.
9
At this point, the thruster is allowed to cool, and the primary and trigger capacitors are recharged.
If continued thrust is required, the cycle will be repeated as quickly as the capacitors can be
charged. In some stationkeeping and attitude control applications, however, reduced pulse rates or
even single-pulse operation are required.
2.2 Typical Parameters
More than half a dozen distinct PPT designs of the classic type described above, have been
developed for flight or advanced laboratory testing. Most have very similar design and
performance parameters, and the LES 8/9 model described below can be considered
representative of the classic PPT design
22
. The LES 8/9 thruster was developed for the USAF
Lincoln Experimental Satellite in the 1970s, but was never flown. Surplus LES 8/9 flight units
have been made available for extensive laboratory testing, and the closely-related XPPT-1 thruster
was designed to duplicate the behavior of the LES 8/9 without using expensive flight-qualified
hardware. This has allowed a great deal of data to be amassed for a common PPT design, which
can serve as a basis for comparison when novel PPT configurations are proposed.
The LES 8/9 is a dual thruster, with two distinct electrode/spark plug/propellant assemblies. Only
one thrust assembly is used at a time. The purpose of this arrangement is to allow a single unit to
provide both stationkeeping thrust and attitude-control torque as needed. The laboratory XPPT-1
has only a single thrust assembly, of identical geometry to that of the LES 8/9

10

Figure 2.2 – The LES 8/9 PPT
The LES 8/9 uses a propellant bar of 2.54cm x 2.54cm face size, and the exposed electrode
surfaces are also of 2.54cm x 2.54cm size. The separation between electrodes is slightly less than
2.54cm, as a step is machined into the electrode face to maintain the propellant bar in its proper
position. Two spark plugs (one only in the XPPT-1) are mounted flush with the cathode surface.
The main discharge capacitor of the LES 8/9 has a capacitance of 17 μF and is typically operated
at 1540 V. The overall discharge circuit, assuming negligible contribution from the plasma arc,
has an inductance of 0.85 μH. The charging circuit will support a 1 Hz pulse rate with an input
power of 25.5 Watts. From these values the following parameters can be derived:
E
d
= ½ C
d
V
d
2
= 20 J (2.1)
F
r
= (C
d
L
d
)
½
= 262 kHz (2.2)
The trigger circuit of the LES 8/9 is more difficult to characterize, because some of the
components of the flight units are unlabeled, undocumented, and potted in epoxy. The supply of
11
LES 8/9 flight units is not sufficient to justify potentially destructive evaluation of component
specifications, so only estimates are available. A schematic of the best current understanding of
the LES 8/9 trigger circuit is given below
Figure 2.3 – The LES 8/9 Trigger Circuit
Of particular relevance are the estimated output voltage of 1800-3000 V and discharge energy of
0.36 – 1.0 Joules.
The LES 8/9 produces a nominal thrust of 300 μN at a specific impulse of 1000 s. This gives
η
d
= 0.5 T g I
sp
/ E
d
F
d
= 7.35% (2.3)
η
tot
= 0.5 T g I
sp
/ P
tot
= 5.75% (2.4)
The LES 8/9 has also been tested in a high-energy mode with E
d
=80J, F
d
=0.25 Hz, T=0.375 mN,
Isp = 1450s, and η
t
= 0.104, though operation in this mode exceeds the design margin for the
system and would not be done in flight. At the other extreme, the thruster can be used to deliver a
single pulse at reduced energy, for a measured I
bit
of 237 μN-s. This is relevant when the system
is used to provide stationkeeping or attitude control force for spacecraft with extremely precise
requirements in that area, e.g. components of a multisatellite interferometry system.
0.01 uF
TTL Trig
+600V - 1 kV
2.0 uF
2 Ω
Spark Gap
12
Finally, the LES 8/9 has a total mass of 7.33 kg, of which 0.75 kg is Teflon propellant. The
demonstrated life of the system is 3.4x10
7
discharges, or slightly less than 9500 hours of
continuous operation.
2.3 The Triggered Micro PPT
As mentioned earlier, most traditional PPT designs have been extremely similar to the LES 8/9.
The only notable exception is the Micro PPT. The μPPT, in its original form, was developed by
Dr. Gregory Spanjers of AFRL in 1998
19
, and represents an attempt to produce a pulsed plasma
thruster of absolute minimum size and complexity. Aside from a reduction in scaling, three
differences from the classic PPT design are observed. First, there is no separate spark plug and
trigger circuit. Because of the reduced scale of the μPPT, it can serve as its own spark plug; self-
breakdown between the electrode gap will occur at reasonable voltages. Second, the spring-fed
propellant feed mechanism is eliminated; the propellant face is allowed to recess into the thruster
as propellant is consumed. Finally, the electrode geometry is coaxial, rather than rectilinear.
Although this is expected to result in a modest performance benefit, for reasons that will be
explained later, the geometry was in fact chosen because it allows the use of commercial semi-
rigid coaxial cable for the μPPT electrode and propellant assembly.
A photograph and a schematic of a self-breakdown μPPT are given below:

Figure 2.4 – The Triggered μPPT
13

Figure 2.5 – Triggered μPPT Schematic

The astute reader may note that the schematic of the μPPT discharge circuit is identical to that of
the LES 8/9 PPT trigger circuit. The discharge voltage and energy appropriate for a μPPT are
similar to those required for the spark plug of a full-sized PPT. This allows the use of a common,
flight-qualified hardware design for both systems. The only other component of the triggered PPT
is a coaxial propellant/electrode module, which for laboratory-model thruster is a length of semi-
rigid coaxial cable. Furthermore, with the addition of a switching box, a single pulsed discharge
circuit could support multiple thrusters on the same spacecraft – a full scale PPT for main
propulsion, and several clusters of μPPTs for attitude control and precision stationkeeping, for
example.
The μPPT thus represents a minimum-complexity PPT system. No separate trigger circuit is
required, the one moving part of the traditional PPT design is removed, and if the host spacecraft
already incorporates a full-sized PPT its trigger circuit can serve as the discharge circuit of any
μPPT on the same spacecraft. Only a switchbox, additional high-voltage cable, and the actual
propellant/electrode assemblies would be required to add a multi-axis μPPT system for precision
stationkeeping and attitude control.
Propellant
Coaxial
+1 kV
2 μF
Trigger
14
This implementation of the μPPT is, however, severely limited in several respects. The low
discharge energy means that even at high pulse rates (the LES 8/9 trigger circuit in isolation can
operate at ~10 Hz) thrust will be extremely low. The low discharge voltage, and the requirement
for self-breakdown across the propellant face, means that the diameter of the propellant bar must
be very small – typically 3.5 mm or less. Finally, the length of the propellant bar must also be
rather small, less than ten centimeters, due to the viscous losses that would ensue when the
propellant has receded the full length of the propellant/electrode assembly and the exhaust is
required to flow through a long, narrow cylindrical electrode before escaping. The short, narrow
propellant bar allows for only a 1-2 grams of propellant, and thus a relatively short thruster life.
Performance of the triggered μPPT is, unfortunately, very hard to measure. The expected
maximum thrust at 1 J discharge energy and 10 Hz pulse rate is only 100-150 μN, whereas the
most precise thrust stand available at AFRL during the period of active research on this
configuration had an error bar of ~200 μN under ideal conditions. On one occasion, a triggered
μPPT was operated at twice the normal discharge energy, and did produce a measured thrust of
250 μN +/- 200 μN. This is not a useful quantitative measurement, it serves only to verify that
estimates of μPPT performance are correct to within an order of magnitude.
2.4 Performance Comparison and Implications.
Table 2.1, below, compares the operational performance of the LES 8/9 PPT, the triggered μPPT,
and a Busek BHT-200 Hall Effect Thruster. The latter is a 200-watt steady-state plasma thruster
developed specifically for small-satellite applications and thus filling the same niche as the LES
8/9 PPT.
15
AFRL μPPT LES 8/9 PPT BHT-200
Thrust, mN 0.15 0.30 11.5
Isp, s 800 1000 1570
Power, W 10 25.5 200
Efficiency, % 5.8% 5.75% 42.7%
Min I
bit
, μN-s 15 237 1750
Dry Mass, kg <1.0 6.6 6.1
Life, hours <100 9500 1500

Table 2.1 – Performance Comparison

It is immediately apparent that the BHT-200 Hall Effect Thruster offers substantially greater
performance than a conventional PPT of the LES 8/9 sort. Substantial increases in PPT thrust and
efficiency would be required for a PPT to be a serious competitor. The PPT does have a
substantially greater operational life, but during the whole of that life will deliver less propulsive
impulse than the BHT-200. Only if precision stationkeeping performance, as reflected by the
minimum I
bit
value, were the overriding concern, would the PPT be at all appropriate.
If the conventional PPT would seem to have been displaced from all but the most highly
specialized applications by the more modern and better-performing Hall Effect Thrusters, the
μPPT may still have a role. It offers an I
bit
smaller even than the conventional PPT, for
applications where that is relevant. Perhaps more importantly, it offers a complete propulsion
system in a ~1kg package. If a spacecraft does not have 5-10 kg in its mass budget for propulsion,
or if that 5-10 kg must support multiple thruster systems, then neither the LES 8/9 nor the BHT-
200 can meet the requirement. And the BHT-200 represents an effort to produce a minimum-
mass Hall thruster in much the same way that the μPPT represents a minimum-mass PPT; there is
unlikely to be a <1kg Hall Thruster in the foreseeable future. Indeed, the only currently available
propulsion systems in the <1kg range are cold-gas thrusters of exceedingly limited performance.
16
However, with the conventional PPT apparently displaced by the low-power Hall thruster, it is no
longer reasonable to believe that full-sized PPT trigger circuits will be available to perform double
duty as μPPT discharge circuits. It is thus highly desirable to reduce the mass and complexity of
the μPPT power electronics. The short operational life of the μPPT is also a serious concern; even
when operated at a low duty cycle for stationkeeping/ACS, it will probably be necessary to
increase the propellant supply beyond the 1-2 grams allowed by the geometric constraints
mentioned earlier.

17

Chapter III

Mission Applications

The list of spacecraft electric propulsion systems proposed and developed at least to the level of
laboratory models includes the resistojet, the arcjet, the microwave electrothermal thruster, the ion
thruster, the Hall effect thruster, the colloid thruster, field emission electrostatic propulsion, the
pulsed plasma thruster, the laser plasma thruster, the magnetoplasmadynamic thruster, the pulsed
inductive thruster, the variable specific impulse MPD rocket, the helicon thruster, and the
electrodynamic tether. The list of electric propulsions actually flown on operational spacecraft, or
likely to be flown on operational spacecraft in the next ten years, is limited to the resistojet, the
arcjet, the ion thruster, the Hall effect thruster, and the PPT.
That the former list is so much larger than the latter, suggests that the world is already quite
oversupplied with laboratory model advanced spacecraft propulsion systems. Adding one more to
the list, would not constitute a meaningful advance in the state of the art, or a productive use of
anyone’s time. Only if a potential new propulsion system is more suitable for some actual space
flight application, is it worth developing. And as shown earlier, existing propulsion systems offer
superior performance to any PPT for most purposes.
The one likely application for the PPT and/or the μPPT, would be to provide propulsion for very
small spacecraft. The minimum mass for a conventional spacecraft propulsion system, chemical
or electric, is approximately 5-10 kg. A typical satellite will devote 2-6% of its total mass to
propulsion systems
23
This suggests that conventional propulsion systems are poorly suited for use
18
by spacecraft of less than 100-500 kg total mass. For spacecraft of substantially less than 100 kg,
conventional propulsion systems are completely inappropriate.
This has not posed a problem in the past, because most operational spacecraft had masses of well
over 500 kg, and the few very small spacecraft launched had missions (usually experimental) of
very limited scope. Correspondingly little in the way of propulsive capability was required;
occasionally a cold-gas thruster set, usually not even that. However, the past decade has seen a
substantial increase in the number of very small spacecraft flown. In 1999, 10 spacecraft of <100
kg total mass were launched. By 2000, that number had increased to 21, and more recently to 75
in 2005
9
. Furthermore, these spacecraft, commonly referred to as microsatellites, are now
commonly proposed for operational, rather than simply experimental, missions, with substantially
more ambitious goals. They are still usually devoid of propulsion systems beyond cold-gas
thrusters, but they are approaching the limits of what can be accomplished within that limit and
there is interest in developing a micropropulsion capability for microsatellite applications.
In 1998, the USAF began the TechSat 21 program to develop and fly proof-of-concept
microsatellites
16
. While primarily a technology development program, the nominal mission to be
demonstrated was distributed-aperture space-based radar. A cluster of microsatellites would
operate in close formation, ~100 meters average separation, with each serving as a single
transmit/receive element of a very large phased-array radar antenna capable of tracking ground
and air targets from space. This architecture is expected to provide greater performance at lower
cost than the use of a single, monolithic antenna.
In order to perform this mission, microsatellite capability will have to be greatly extended from the
state of the art in almost every respect. This is of course quite appropriate for a technology
19
development program. Of particular relevance here, is that the TechSat 21 satellites will require a
substantial propulsive capability. Unlike previous microsatellite systems, the TS21 spacecraft will
have to maintain a precisely controlled formation for a period of a year or so even in an
experimental system, and perhaps a decade for an operational system. This is well beyond what
can be achieved with a cold gas thruster system; some form of advanced micropropulsion will be
necessary for this mission.
Many other microsatellite programs exist, in various states of development. We will
consider the TechSat 21 representative of the class; it is the USAF’s flagship microsat
program, and its propulsion requirements are near the high end of those expected for
microsatellites as a class. If the PPT and/or the μPPT is suitable for the TechSat 21 mission,
it is probably suitable for microsatellite propulsion in general. During 1998/1999, a survey
of potential micropropulsion technologies was performed at AFRL to resolve this issue.

3.1 TechSat 21 Spacecraft Design
At the time of this study the TechSat 21 spacecraft were in conceptual design phase. The
proposed design, shown in Fig. 3.1, collapses into a 0.3 cubic meter volume for launch, then
deploys a 7-meter boom and 2.5 meter antenna on orbit. As shown in Table 1, the total mass
of the spacecraft is ~135 kg, of which ~10 kg is available for the propulsion system. Given
the likelihood of weight growth, we will size the propulsion system for a 150 kg spacecraft.
For attitude control purposes, we assume moments of inertia I
zz
= 50 kg-m
2
and I
xx
= I
yy
=
1000 kg-m
2
.

20

Figure 3.1 – TechSat 21 Spacecraft

Approximately 350 W of electric power is produced by solar panels on the boom section,
almost all of which is available for the propulsion system during the maneuver phases of the
mission. No estimate is available for the amount of power available for stationkeeping
propulsion when the propulsion system must compete with the radar transmitter for available
power, but as the stationkeeping thrust requirement is small compared to the maneuver
requirement, any propulsion system capable of performing the maneuver mission with < 350
W of power will almost certainly be able to perform the stationkeeping mission with
minimal impact.

2.4m
7 m
Deployed
TRAM Antenna Modules,
2.3 kg each
Multi-Functional
Bus Module, 24 kg
Thin film
solar array,
1.0kW, 15 kg
Battery module, 16 kg
Propulsion module, 9 kg
Deployment/
Gravity gradient
boom
21
The proposed TechSat 21 spacecraft is passively stabilized by gravity gradient using the
extended solar array boom, with magnetic torquers for attitude control. No momentum
wheels or other precise three-axis attitude control system is planned, and the magnetic
torquers provide only coarse attitude control with +/- 10 degree accuracy. This may be
insufficient for propulsive maneuvers. The gravity gradient system, of course, provides no
yaw control capability and limits the spacecraft to small deviations about a single pitch/roll
orientation.

The propulsion requirement specified by the TechSat21 program for the initial flight
experiment is as follows:
• Total ΔV of 70 m/s
• Formation maneuver ΔV of 30 m/s over 30-day maneuver period
• Stationkeeping ΔV of 40 m/s over 1-year life in ~600 km orbit
• Total of 5-10,000 propulsive maneuvers
• Minimum impulse bit of 2 mN-s

A series of formations with separation distances from 5 meters to 5 kilometers are to be
tested during the first TechSat 21 flight experiment; the 30 m/s ΔV budgeted for
maneuvering allows shifting between formations. A nominal thirty days are devoted to such
maneuvers, which corresponds to a minimum thrust level of 2 mN. However, higher thrusts
and shorter maneuvering periods are obviously preferable. Conversely, it may be possible to
accept lower thrust and longer maneuver time if substantial savings in propulsion system
mass, cost, and complexity result.

22
An operational system is expected to have rather different propulsive requirements, with a
longer mission duration and higher orbit altitude. This would also be true of any operational
SBR system derived from TechSat 21 experience. Preliminary estimates for the TechSat 21
program, appropriate for operational systems, call for 390 m/s total ΔV broken down as
follows:
• Orbit raising ΔV of 50 m/s
• Drag makeup ΔV of 20 m/s over 10-year life
• Stationkeeping ΔV of 200 m/s over 10-year life
• Deorbit ΔV of 120 m/s at end of life.

The first mission of the propulsion system is to carefully increase the orbital altitude to that
of the planned constellation and insert the satellite in phase with the rest of the formation.
This will dominate the thrust requirement, and again we allow a nominal 30 days for the
maneuver.

Because of the substantially higher ΔV requirement, an operational system, may call for a
different propulsion system than the initial flight experiment. This study is focused on
propulsion system selection for the experimental spacecraft; however, in recognition of the
fact that a common propulsion system for both missions will simplify spacecraft design, the
suitability of various propulsion options to the later flight will be considered.

Stationkeeping and formation forming maneuvers are expected to be performed once or
twice per orbit, for a total of 5-10,000 such maneuvers over the course of the mission. This
corresponds to an average thrust impulse of 1-2 N-s per maneuver. In certain extreme cases
23
involving spacecraft maneuvering in close proximity (~5 meters separation), impulse bits as
low as 2 mN-s may be required.

All of these maneuvers are expected to be confined to the spacecraft X-Y plane, so only yaw
steering will be required of the propulsion system. There are three potential approaches to
this requirement. First, the coarse yaw steering capability provided by the spacecraft’s
magnetic torque rods can be utilized, albeit with some sacrifice of precision maneuvering
capability. Second, multiple propulsive thrusters with canted nozzles can be used to provide
yaw steering capability. Finally, a dedicated micropropulsion system can be used for yaw
pointing and precision maneuvering.

The thrust requirement for such an ACS thruster system can be determined by considering
the worst-case scenario. With maneuvers occurring twice per orbit, it could be necessary to
rotate the spacecraft by a full 180 degrees in one-half of an orbit period. Given an orbit
period of ~6000 seconds and a spacecraft moment inertia I
ZZ
= 50 kg-m
2
, this requires a
torque of 6.8x10
-5
N-m. The configuration of the spacecraft favors ACS thrusters with a
moment arm of 0.6 meters, for 1.13x10
-3
Newtons of available thrust. If this is to be
provided by a single pair of propulsive thrusters with nozzles canted at 30 degrees, a thrust
level of ~200 μN per thruster will be required. If by three opposed pairs of dedicated ACS
microthrusters, 38μN per microthruster will be required.

In most cases, of course, the spacecraft will not have to perform a full 180 degree rotation in
half an orbit period. Maneuvers are more likely to be performed once per orbit, and to
require consistent and predictable pointing. In particular, approximately 90% of the
24
stationkeeping mission consists of ascending-node correction for J
2
perturbation and drag
makeup, both of which can be performed with burns in a single orientation once per orbit.
Only the remaining 10% of the stationkeeping requriement would call for yaw pointing,
presumably of random character. Over a one-year mission with 5000 stationkeeping
maneuvers, this would require a total of 32 N-m-s of torque impulse.

The formation maneuvers are somewhat less predictable, and will be assumed to be evenly
divided between consistent and random orientation. This greater yaw pointing requirement
will call for 113 N-m-s of torque impulse over a one-year mission. Also, it may be
necessary to provide yaw stabilization against disturbance torques. The worst case in this
regard would be an aerodynamic torque due to a cg misalignment of 3 cm (5% of vehicle
characteristic width). This would produce a disturbance torque of ~0.75 μN-m, requiring 24
N-m-s of torque impulse per year for stabilization.

These combined attitude control requirements come to 169 N-m of torque impulse for the
one-year mission. If provided by the main propulsive thrusters, again with canted nozzles at
30 degrees, this corresponds to 560 N-s of total impulse, or 3.75 m/s of ΔV for a 150-kg
spacecraft. Dedicated ACS microthrusters would have to provide only 281 N-s of total
impulse, 47 N-s for each of six thruster units, corresponding to 1.9 m/s of ΔV. With three
ACS thruster pairs to chose from, it will be possible to apply most (~81%) of this ΔV to the
stationkeeping as well as yaw steering requirements, which is not the case if only a single
pair of main propulsive thrusters must perform the yaw pointing requirement. Furthermore,
canting the nozzles of the main propulsive thrusters by 30 degrees reduces their efficiency
for stationkeeping and formation maneuvers by 13%.
25

Thus, assigning the yaw steering requirement to the main propulsive thrusters requires the addition
of a second such thruster, and increases the net ΔV requirement to 85 m/s. If dedicated
microthrusters are used, three pairs of thrusters each of 38 μN thrust and 47 N-s impulse capability
will be required, and the ΔV requirement for the main thrusters is reduced to 68.5 m/s. If only the
magnetic torque rods are used, there is no cost to the propulsion system, but precision pointing
(and thus maneuvering) are sacrificed.
3.2 Micropropulsion Options
Traditionally, on-orbit propulsion for spacecraft has been provided by chemical rockets.
Solid-propellant rockets are the simplest propulsion system available, but cannot be used for
stationkeeping due to the inability to repeatedly turn them on and off. Liquid propellant
rockets are more versatile, and offer greater performance, albeit at the cost of increased
complexity. Cold-gas thrusters offer sufficient versatility for stationkeeping in a simpler
package than liquid rockets, but have extremely limited performance. Small liquid and cold-
gas systems are considered for the TechSat21 mission in their conventional forms, with the
liquid-propellant system serving as a baseline against which other candidate propulsion
systems will be measured. In addition, a novel form of miniaturized solid rocket propulsion
system, the digital MEMS thruster, is suitable for spacecraft in the 100kg and under size
class and is included in the analysis.

Unfortunately, no chemical propulsion system can offer a specific impulse greater than ~550
seconds, and systems suitable for use on small spacecraft are limited to approximately 220
seconds. This is adequate for the initial flight experiment, but the low specific impulse is likely to
26
result in unacceptably high propellant mass values for more demanding operational missions.
However, electric propulsion systems are available which offer much higher specific impulse at
the expense of a substantial electric power requirement. As the proposed TechSat21 spacecraft
has a substantial solar-electric power capability for the radar payload, while the mass budget for
the propulsion system is rather tight, electric propulsion may be preferable for the TechSat21
mission.
Two general categories of electrical propulsion system will be considered for the TechSat21
mission, electrostatic propulsion and electromagnetic propulsion. The former is characterized by
the generation of charged particles, usually heavy ions, which are accelerated by an applied
potential to velocities in excess of ten kilometers per second to produce thrust. One type of
electrostatic system is the ion thruster, in which the applied potential is provided by a series of
charged grids, with the ions provided by a separate discharge chamber. Conventional ion thruster
designs do not readily scale down to the size and power levels required for TechSat21 due to
discharge chamber physics, but novel ion sources currently under development may allow micro-
scale ion thrusters (e.g field emission electrostatic propulsion (FEEP), micro colloid thruster) to
meet the TechSat21 mission requirements. Another type of electrostatic propulsion system is the
Hall-effect thruster (HET), also known as the stationary plasma thruster (SPT). In this design, the
accelerating potential is provided by forcing a discharge current through a transverse magnetic
field. There are several demonstrated SPT designs suitable for the TechSat21 mission.
Electromagnetic propulsion involves the acceleration of a current-carrying plasma by an applied
or self-generated magnetic field, rather than by an electrostatic potential. The requirement for a
strong magnetic field limits steady-state electromagnetic thrusters to extremely high power levels,
which is unacceptable for the TechSat21 mission. However, pulsed operation allows for
27
arbitrarily low average powers, with the high peak power requirement being met by a capacitive
discharge, and the pulsed plasma thruster (PPT) is a leading candidate for the TechSat21 mission.
The PPT’s use of an inert solid propellant with no moving parts is a particularly desirable feature
for a system intended for use in a small, low-cost spacecraft.
Several interesting microthruster concepts are not addressed in this analysis, such as the
Vaporizing Liquid Microthruster (JPL), a Free-Molecular Micro Resistojet (AFRL), and other
electrothermal devices. The high thrust of these devices makes them attractive for some
microsatellite missions, however their low specific impulse results in an excessive propellant mass
for missions with substantial total ΔV requirements, such as TechSat 21
A. Chemical Micropropulsion
Liquid-propellant rockets have been the standard for on-orbit propulsion throughout the history of
space travel, and no introduction will be given. Due to the limited mass budget of TechSat21, and
the implied requirement for simplicity, only monopropellant systems are considered, with the
baseline monopropellant being hydrazine. The numerous handling difficulties of hydrazine
notwithstanding, hydrazine monopropellant thrusters and propellant feed systems are mature,
commercial products and can be integrated with the TechSat21 spacecraft with little difficulty.
For the purposes of this analysis, the Primex MR-111 and MR-103M thrusters were specified,
though numerous other manufacturers offer equally suitable systems. , These represent the
smallest commercially available chemical thrusters, and while somewhat larger than optimal for
TechSat21 are still reasonable for the mission.
While hydrazine monopropellant thrusters may offer insufficient specific impulse for this
application, other monopropellant options are available. The Air Force Research Laboratory is
28
currently developing a series of monopropellants based on hydroxyl ammonium nitrate (HAN),
which promise to deliver up to 25% greater specific impulse than hydrazine. The high
combustion temperature of these propellants requires the use of new materials for thruster
construction, and there are also concerns regarding the long-term stability of the new propellants.
Nontheless, we will consider an AFRL advanced monopropellant formulation, AF-315A, for the
TechSat21 mission. Thrusters designed for use with AF-315A will be assumed 25% heavier than
comparable hydrazine thrusters due to design requirements imposed by the high chamber
temperature.
Cold-gas thrusters are the simplest throttleable thruster available, and thus the simplest propulsion
system suitable for the stationkeeping and attitude control requirements. Unfortunately, the
combination of extremely low specific impulse and heavy, high-pressure propellant tanks results
in unacceptably high total propulsion system mass values even though the thrusters themselves
can be quite small. Cold gas thrusters are wholly unsuited for an operational microsatellite, and
are only marginally capable of performing TechSat 21 experimental mission. However, they do
have the advantage of using an inert, gaseous propellant and thus do not contaminate exposed
spacecraft surfaces. Because of the threat of mutual contamination when spacecraft operate
propulsion systems in close formation, the combination of cold-gas thrusters for attitude control
and precision stationkeeping and a similarly non-contaminating main propulsion system will be
considered, with the Moog 58-102 thruster baselined for analysis.
While all of the aforementioned chemical propulsion systems can be obtained in sizes suitable for
the TechSat21 mission, they begin to suffer from scaling effects at that level. Chemical
propulsion would be largely unsuitable in future microsatellite missions with mass budgets an
order of magnitude smaller. To meet the microsatellite requirement, several institutions have
29
proposed the Digital MEMS (Micro ElectroMechanical System) thruster. This device uses
semiconductor manufacturing techniques to etch thousands of extremely small (~500 mm)
cavities and nozzles into a silicon wafer. Each cavity is filled with a propellant charge and serves
as a one-shot microthruster at need. The specific impulse and propellant mass fraction suffer in
comparison with conventional chemical rockets, but the ability to scale down to extremely small
sizes compared to conventional systems is desirable for the microsatellite application. The
availability of small discrete thrust impulses is particularly advantageous for stationkeeping and
attitude control. Both TRW and Honeywell presently have programs to fabricate digital MEMS
thrusters and have tested the necessary igniter arrays. While there are still substantial technical
challenges associated with the concept it will be considered as an option for TechSat21, using the
performance estimated by TRW and Honeywell.
B. Electromagnetic Micropropulsion
The only electromagnetic micropropulsion systems considered here are the PPT and the μPPT,
which have already been described. A hypothetical PPT corresponding to the goals of the AFRL
PPT development effort at the time of the study, and a self-breakdown μPPT sharing the same
trigger circuit as the full-sized PPT, are considered. The performance figures for the full-sized
PPT represent a substantial increase over the LES 8/9 type state of the art system, and would
require a substantial research and development effort to achieve.
For applications on space based radar microsatellites, the radiated EMI from the propulsion
system poses a serious concern. For the PPT, GHz radiation from the spark ignitor discharge can
interfere with the primary transceiver frequencies. EMI radiation from the main PPT discharge
can interfere with the radar frequency shifts in the MHz range. Although it is still a topic of
30
current research, it is believed that the co-axial geometry of the μPPT and of proposed next-
generation PPTs will better confine the EM radiation.
C. Electrostatic Micropropulsion
The high specific impulse, high efficiency, and modest mass have made Hall effect thrusters the
electric propulsion system of choice for many future missions at power levels of 1 kW or above.
One recent development in the field is a reduced-scale 200W Hall thruster applicable to the
TechSat21 mission as well. HETs were developed in the former Soviet Union during the 1960s
and 1970s, and have been flown in over a hundred successful missions. The relatively high mass
of even the smallest HETs, though, renders them marginal at best for stationkeeping and attitude
control despite the potential savings associated with the use of a common power processing unit
and propellant feed system. For purposes of this analysis the BHT-200 Hall thruster developed by
Busek Corp. is considered for primary propulsion.
Another electrostatic thruster proposal under consideration is Field-Effect Electrostatic Propulsion
(FEEP). This is an ion thruster using a field emission ion source, in the form of a narrow slit
anode through which cesium propellant is passed and ionized by the geometrically-enhanced
electric field. This offers a more compact and efficient ion source than the traditional electron-
bombardment ionization chambers used with ion thrusters, allowing the extension of electrostatic
propulsion to smaller spacecraft than previously possible. FEEP systems have been demonstrated
in the laboratory and to some extent in flight experiments. These systems offer extremely high
specific impulse values at reasonable efficiency, but specific power and thrust is low. It may,
therefore, be necessary to relax the 30-day maneuver time requirement for a FEEP-based
31
propulsion system. For the purpose of this analysis, several combinations of the Centrospazio
120-watt FEEP thruster with complimentary ACS thrusters will be considered.
The basic physical principles of the FEEP thruster can be scaled down to the true micropropulsion
regime, resulting in the micro colloid thruster. As with the digital chemical microthruster
described earlier, this system consists of a large number of discrete thrusters micromachined into a
silicon wafer. The thruster elements consist of microvolcano field-emission ion sources, in which
a propellant is fed through a small, sharp needle, which serves as an anode similar to the Spindt-
type microcathode. The field enhancement associated with the sharp tip of the needle results in
the emission of charged micron-scale droplets of propellant. The droplets are accelerated by an
applied electric field and neutralized by an external electron source, or perhaps by a parallel
microthruster element operating at the opposite polarity. Systems proposed by Phrasor Scientific
and MIT use doped glycerol propellants and are predicted to achieve specific impulse values of
order 1000 seconds at reasonably high efficiency. As yet, only limited progress has been made in
fabricating thruster subassemblies and testing representative components, and no actual thruster
has been constructed or tested, so there would be a high degree of technical risk associated with
the use of micro colloid thrusters in TechSat21.
D. Electric Power Processing
It should be noted that for all of the electric propulsion systems described, dedicated power
processing hardware is required. The TechSat21 spacecraft will have ample electric power
available from the solar array boom, but the main bus can be expected to operate at less than
100V, while the various electric thrusters require anywhere from 300 V to 10 kV. Also, few of
the systems have been tested in a simple direct-drive configuration. Traditionally, voltage- or
32
current-regulated switching power supplies are used, which tend to outweigh the thrusters they
drive by a factor of two or three. We assume that such conventional power supplies will be used
with the Hall thruster, but for the various micropropulsion systems a monolithic solid-state DC-
DC converter seems a more reasonable choice. A Micro-PPT has been operated using such a
system at AFRL, and it will be assumed that similar systems can be used with the colloidal and
MFIT systems. With PPTs, a trigger unit and a discharge capacitor must also be provided, and are
included in the system weight estimate. Notwithstanding the requirement for a power processing
unit, such systems are rightly considered separately from the thrusters themselves, as a single PPU
can serve a substantial number of distinct thrusters in stationkeeping or attitude control operation.
In the case of the PPT, it is fortuitously possible for the PPT trigger circuit to also serve as the
entire power-processing unit for an associated Micro-PPT system.
Type
Isp η Mass Thrust Power
Cold Gas Thruster 75 s 95+% 0.01 kg 5 mN N/A
Solid Rocket Motor 185 s 90+% 1.60 kg 100+ N N/A
Digital MEMS 200 s ~75% 0.04 kg 50 mN N/A
Hydrazine Monopropellant 220 s 95+% 0.32 kg 2 N N/A
Advanced Monopropellant 290 s 95+% 0.40 kg 2 N N/A
Colloidal Microthruster 500-1500 s ~50% 0.08 kg* 20 μΝ 0.2 W
Pulsed Plasma Thruster 1000 s ~10% 0.25 kg* 700 μN 40 W
μPPT 1000 s ~5% 0.12 kg* 40 μN 4 W
Hall Thruster 1500 s ~35% 1.00 kg* 10 mN 200 W
FEEP 8,000 s ~25% 3.50 kg* 800 μN 120 W
MFIT 15,000 s ~90% 0.15 kg* 1.5 mN 300 W
* Values for EP systems do not include power processing unit
η = total system efficiency (thrust power output to chemical or electrical power input)

Table 3.1 – Mircropropulsion Options
A comparison of the thruster proposals described above is given in Table 3.1. As can be seen,
electric propulsion systems generally offer specific impulse values of 1000-1500 seconds and
chemical systems approximately 200 seconds. Electric propulsion is thus quite likely to offer
33
lower overall system mass, presuming the thruster and PPU masses can be kept to acceptable
levels. Thrusters can also be divided into micro- and macro scales, with the macro thrusters being
small versions of conventional spacecraft propulsion systems and the microthrusters developed
specifically for microsatellite applications, often using semiconductor-style, commonly referred to
as MEMS, fabrication. The macrothrusters, with masses of order 1 kg, are generally suitable for
the main propulsion application, but are too heavy to be used in the numbers required for the
stationkeeping/ACS application. The microthrusters, at approximately 100 grams, can meet the
stationkeeping/ACS requirement but lack the thrust needed for main propulsion of the 100kg
TechSat21 vehicle unless used in clusters. Some combination of the two is likely required.
3.3 Analysis
Given the extremely tight mass budget set for the TechSat21 spacecraft, any comparison of
propulsion options must center on predictions of propulsion system mass. To address this issue,
detailed mass estimates for propulsion systems using the various proposed technologies were
constructed. A total of 26 propulsion options were considered, with each of the potential main
propulsion systems matched with one or more compatible stationkeeping systems. Mass
estimates for each system were broken down into five categories - thruster, PPU, propellant,
propellant feed, and miscellaneous - with one or more line items in each category as appropriate.
Separate evaluations were made for the requirements of flight experiment and follow-on
operational missions, due to the different ΔV and propellant requirements.
The thruster category includes the main propulsive thruster or thrusters and the stationkeeping
thrusters. The size and/or number of main thrusters was set by the ~2 mN thrust requirement for a
total thirty-day maneuver period, except in the case of the FEEP system where a relaxed 60-day
34
requirement was allowed due to its low thrust-to-power ratio. In some cases, such as the Hall
thruster or any of the chemical systems, a single thruster of the smallest reasonable size provided a
much shorter maneuver period. For attitude control and stationkeeping, a total of six ACS
thruster elements with a minimum thrust of 40 mN were required. Any special mounting
hardware required was also included in this category.
For electric propulsion systems, a high-voltage DC power-processing unit is invariably required,
as previously described, and the use of semiconductor DC-DC conversion has been postulated in
most cases. Also included in this category are any necessary high-voltage or high-current cables.
In the case of PPTs, an energy storage capacitor and a pulse trigger generator are also required. If
the main and ACS systems incorporate different electric propulsion technologies, separate power
processing systems are generally required. The principal exception to this rule is the ability of a
PPT trigger pulse generator to serve as the entire power processing system of a μPPT attitude
control system although additional switching is required.
Sufficient propellant was provided to meet the specified ΔV requirements of the flight experiment
and operational missions, as described earlier. In some cases, it was deemed advantageous to
provide separate propellant storage for each ACS thruster element rather than a feed system from
a central tank. For these cases the requirement is set at 100 N-s total impulse per element for the
flight experiment and 1 kN-s for the operational mission to account for possible non-uniform
propellant usage by the ACS. The propellant feed system includes tankage for main and ACS
propellant, feed lines, valves, and flow control systems.
Ten percent of the propulsion system net dry mass is specified for structures and general mounting
hardware. An additional five percent is specified for control systems and wiring harnesses using
35
standard spacecraft design practice. This is exclusive of any high-voltage distribution system
incorporated in the PPU category. Finally, a fifteen percent margin is set aside for unexpected
system growth. This total of 30% of net dry mass constitutes the miscellaneous category.
For each of the enumerated items, commercial off-the-shelf hardware was specified whenever
possible, preferably space-qualified but in the case of some PPU or propellant feed system
components, ground or aviation hardware meeting relevant military specifications was used as a
baseline. The intention is to reliably estimate the mass of a flight system rather than to actually
design such a system. In some cases, commercial systems of different power levels were scaled
linearly over a modest range to meet specific TechSat21 requirements. For experimental thruster
concepts, flight-like laboratory test hardware was considered, and in the case of some technologies
which have not yet reached even the test stage, the best estimates of the authors regarding
developed system weights were used.
Space precludes giving the detailed mass breakdowns for all propulsion system options here,
though a representative sample is given in Table 3.2. Figure 3.2 is a schematic layout of the same
system, indicating the major components. While specific to the all-PPT propulsion option, other
propulsion systems will have a similar configuration. Six ACS thrusters in a trilateral arrangement
are specified for X-Y stationkeeping and yaw control, rather than the traditional eight-thruster
orthogonal arrangement. While this does result in a small (< 13%) reduction in efficiency due to
cosine losses, the ease of integration with the hexagonal TechSat21 bus and the ~25% reduction in
system dry mass due to the reduced number of thruster units more than compensates for non-
orthogonal losses, and leads to the recommendation of the trilateral system for this application.

36
Component Type # Unit Mass Total Mass
Main Thruster CU Aerospace PPT-7 4 420 gm 1680 gm
Propellant Teflon bars 8 165 gm 1320 gm
Capacitor Unison 1 1460 gm 1460 gm
Power Processing Unit Unison 1 2090 gm 2090 gm
Trigger Pulse Unit Unison 1 330 gm 330 gm
ACS Thrusters AFRL μPPT modules 6 60 gm 360 gm
High-Voltage Cable RG-58 or equivalent 2.5 m 40 g/m 100 gm
SCR Switch Modules 6 20 gm 120 gm
Structures & Mounts 10% of Dry Mass 600 gm
Controls & Wiring 5% of Dry Mass 300 gm
Design Margin 15% of Dry Mass 900 gm
Total 9.25 kg
Component Type # Unit Mass Total Mass
Main Thruster CU Aerospace PPT-7 4 420 gm 1680 gm

Table 3.2 – TechSat 21 Sample Propulsion System Mass Breakdown

Figure 3.2 – TechSat 21 Propulsion System Layout

PPT, x4
(Main Propulsion)
Capacitor
Transmission
Line
Micro-PPT
Pulser
SCR, x10
Micro-PPT Clusters, x6
(Attitude Control)
37
Tables 3.3 and 3.4 provide comparative breakdowns of all the concepts included in this study, for
the flight experiment and operational missions, respectively. Most of the chosen propulsion
systems can meet the specified 10 kg propulsion system mass requirement for the flight
experiment. Unfortunately, none of the proposed systems can meet this mass requirement for the
follow-on operational system. A number of electric propulsion systems offer total masses in the
12-15 kg range which, for lack of any reasonable alternatives, must be considered acceptable.
The conventional chemical monopropellant baseline, at more than twice the budgeted mass, is not
a reasonable candidate for the operational mission.

Propulsion Type System Mass Breakdown
Main ACS Thruster PPU Propellant Tankage Misc Total
MFIT MFIT 0.45 kg 2.70 kg 0.10 kg N/A 0.95 kg 4.20 kg
HAN MEMS 0.60 kg N/A 4.30 kg 0.70 kg 0.40 kg 6.00 kg
HAN Magnetic 0.40 kg N/A 5.05 kg 0.80 kg 0.45 kg 6.70 kg
HAN μPPT 0.65 kg 0.55 kg 4.20 kg 0.75 kg 0.60 kg 6.75 kg
HAN Colloid 1.00 kg 0.35 kg 4.20 kg 0.75 kg 0.65 kg 6.95 kg
N
2
H
4
MEMS 0.50 kg N/A 5.65 kg 0.85 kg 0.40 kg 7.40 kg
Colloid Colloid 3.30 kg 0.90 kg 1.75 kg 0.25 kg 1.30 kg 7.50 kg
HAN HAN 1.05 kg N/A 4.80 kg 1.15 kg 0.65 kg 7.65 kg
Hall Magnetic 2.00 kg 2.00 kg 0.75 kg 1.35 kg 1.60 kg 7.70 kg
N
2
H
4
Magnetic 0.30 kg N/A 6.65 kg 0.50 kg 0.45 kg 7.85 kg
Hall MEMS 2.15 kg 2.00 kg 0.90 kg 1.35 kg 1.55 kg 7.95 kg
N
2
H
4
μPPT 0.55 kg 0.55 kg 5.50 kg 0.85 kg 0.60 kg 8.05 kg
N
2
H
4
Colloid 0.90 kg 0.35 kg 5.50 kg 0.85 kg 0.65 kg 8.25 kg
Hall μPPT 2.25 kg 2.55 kg 0.75 kg 1.35 kg 1.85 kg 8.75 kg
PPT MEMS 1.85 kg 3.75 kg 1.45 kg N/A 1.75 kg 8.80 kg
PPT Magnetic 1.70 kg 4.05 kg 1.50 kg N/A 1.70 kg 8.95 kg
Hall Colloid 2.60 kg 2.35 kg 0.75 kg 1.35 kg 1.90 kg 8.95 kg
PPT μPPT 1.90 kg 4.10 kg 1.45 kg N/A 1.80 kg 9.25 kg
N
2
H
4
N
2
H
4
0.95 kg N/A 6.30 kg 1.25 kg 0.65 kg 9.15 kg
PPT Colloid 2.25 kg 4.10 kg 1.55 kg N/A 1.90 kg 9.80 kg
MEMS MEMS 3.55 kg N/A 5.30 kg N/A 2.05 kg 10.90 kg
PPT PPT 3.35 kg 4.30 kg 1.55 kg N/A 2.30 kg 11.50 kg
FEEP Magnetic 7.00 kg 4.10 kg 0.15 kg 0.60 kg 3.50 kg 15.35 kg
FEEP MEMS 7.15 kg 4.10 kg 0.35 kg 0.60 kg 3.55 kg 15.75 kg
FEEP μPPT 7.25 kg 4.65 kg 0.25 kg 0.60 kg 3.75 kg 16.50 kg
FEEP Colloidal 7.60 kg 4.45 kg 0.25 kg 0.60 kg 3.70 kg 16.60 kg

Table 3.3 – TechSat 21 Propulsion System Mass (Flight Experiment)
38
Propulsion Type System Mass Breakdown
Main ACS Thruster PPU Propellant Tankage Misc Total
MFIT MFIT 0.45 kg 2.70 kg 0.40 kg 0.10 kg 0.95 kg 4.60 kg
Colloid Colloid 3.30 kg 0.90 kg 5.00 kg 0.60 kg 1.30 kg 11.10 kg
Hall Magnetic 2.00 kg 2.00 kg 4.50 kg 2.10 kg 1.85 kg 12.45 kg
Hall μPPT 2.45 kg 2.55 kg 4.10 kg 1.95 kg 2.10 kg 13.15 kg
Hall Colloid 2.60 kg 2.35 kg 4.05 kg 2.05 kg 2.10 kg 13.15 kg
PPT MEMS 2.20 kg 3.75 kg 7.50 kg N/A 1.85 kg 15.30 kg
PPT μPPT 2.10 kg 4.10 kg 7.50 kg N/A 1.85 kg 15.50 kg
PPT Colloid 2.25 kg 4.10 kg 7.20 kg 0.10 kg 1.90 kg 15.55 kg
PPT Magnetic 1.70 kg 4.05 kg 8.55 kg N/A 1.70 kg 16.00 kg
Hall MEMS 3.70 kg 2.00 kg 5.95 kg 2.10 kg 2.35 kg 16.10 kg
FEEP Magnetic 7.00 kg 4.10 kg 0.85 kg 0.65 kg 3.50 kg 16.10 kg
FEEP MEMS 7.50 kg 4.10 kg 1.45 kg 0.65 kg 3.75 kg 17.45 kg
FEEP Colloidal 7.60 kg 4.45 kg 1.00 kg 0.75 kg 3.75 kg 17.55 kg
FEEP μPPT 7.25 kg 4.65 kg 1.15 kg 0.65 kg 3.80 kg 17.70 kg
HAN μPPT 0.85 kg 0.55 kg 17.95 kg 0.75 kg 0.65 kg 20.75 kg
HAN Colloid 1.00 kg 0.35 kg 18.00 kg 2.25 kg 1.10 kg 22.70 kg
N
2
H
4
μPPT 1.20 kg 0.55 kg 18.30 kg 2.10 kg 1.15 kg 23.30 kg
HAN HAN 1.05 kg N/A 20.10 kg 2.35 kg 1.00 kg 24.50 kg
HAN MEMS 1.50 kg N/A 20.10 kg 2.10 kg 1.10 kg 24.80 kg
HAN Magnetic 0.40 kg N/A 21.90 kg 2.50 kg 0.85 kg 25.65 kg
N
2
H
4
Colloid 0.90 kg 0.35 kg 23.35 kg 2.70 kg 1.20 kg 28.50 kg
PPT PPT 3.35 kg 4.30 kg 8.15 kg N/A 2.30 kg 29.60 kg
N
2
H
4
N
2
H
4
0.95 kg N/A 25.90 kg 2.90 kg 0.40 kg 30.15 kg
N
2
H
4
MEMS 1.95 kg N/A 25.35 kg 2.65 kg 1.40 kg 31.35 kg
N
2
H
4
Magnetic 0.30 kg N/A 28.15 kg 2.65 kg 1.10 kg 32.20 kg
MEMS MEMS 18.00 kg N/A 27.05 kg N/A 5.40 kg 50.45 kg

Table 3.4 – TechSat 21 Propulsion System Mass (Operational Mission)

Factors other than propulsion system mass must also be considered in comparing TechSat21
propulsion options. In particular, the technical maturity of the various systems is a major concern.
Only the HET, FEEP and chemical monopropellant systems could be constructed using existing
flight-qualified hardware. The PPT and μPPT thruster systems have at least been demonstrated in
the laboratory, and are based on existing flight-qualified systems operating at higher power levels.
Less tested systems such as the digital MEMS and colloidal microthruster are higher risk, and
therefore may suffer increases in mass and cost while developing a flight unit.
39
Also relevant are the maneuver time and power requirement, though it is clear that any of the
proposed systems can achieve acceptable performance in these regards. In particular, with a
power budget of 350 W the TechSat21 propulsion system is likely to be mass-limited rather than
power-limited. Power processing units capable of handling the full available power would be
excessively heavy (5+ kg for the PPU alone). All of the electric propulsion systems found
competitive for TechSat21 operate at power levels of less than two hundred watts.
Finally, the precision of the attitude control system may be an important consideration. Any of the
systems listed can meet the preliminary requirements specified by the TechSat 21 program office,
but increasing emphasis on close-proximity maneuvering is likely to tighten those requirements.
Systems based on a single, large thruster with attitude control provided by magnetic torquers alone
offer only coarse directional control, which may be problematic. Differential throttling of two or
more main thrusters can offer greater precision, but introduces coupling between yaw steering and
X-Y plane translation. Only systems with a dedicated, microthruster-based attitude control system
can provide truly high precision.
Table 3.5 lists these parameters and issues for all studied propulsion system options. These
systems fall into three general categories. The colloidal microthruster would seem to be preferable
on virtually all technical grounds, and is the only system capable of meeting the mass budget for
the operational mission. However, its low technical maturity is problematic for an operational
system and absolutely rules out consideration for a near-term flight experiment mission. The
Hall-μPPT and PPT-μPPT combinations offer total system mass only slightly exceeding the
mission mass budget, and meet all other requirements with a high degree of technical maturity.
Finally, chemical-based systems deliver adequate mass and performance for the demonstration
40
mission, but become excessively massive if called upon to meet the high ΔV requirements of the
operational mission.
Propulsion Type Propulsion Mass

Main

ACS

Demo

Oper
Maneuver Maneuver
Time
Technology
Status
ACS
Precision
MFIT MFIT 4.20 kg 4.60 kg 300 W 60 days Research High
HAN MEMS 6.00 kg 24.80 kg 5 W < 1 day Research High
HAN Magnetic 6.70 kg 25.65 kg 5 W < 1 day Demonstrated Low
HAN μPPT 6.75 kg 20.75 kg 5 W < 1 day Demonstrated High
HAN Colloid 6.95 kg 22.70 kg 5 W < 1 day Research High
N
2
H
4
MEMS 7.40 kg 31.35 kg 5 W < 1 day Research High
Colloid Colloid 7.50 kg 11.10 kg 15 W 30 days Research High
HAN HAN 7.65 kg 24.50 kg 5 W < 1 day Demonstrated Medium
Hall Magnetic 7.70 kg 12.45 kg 200 W 10 days Demonstrated Low
N
2
H
4
Magnetic 7.85 kg 32.30 kg 5 W < 1 day Flight-Ready Low
Hall MEMS 7.95 kg 16.10 kg 200 W 10 days Research High
N
2
H
4
μPPT 8.05 kg 23.30 kg 5 W < 1 day Demonstrated High
N
2
H
4
Colloid 8.25 kg 28.50 kg 5 W < 1 day Research High
Hall μPPT 8.75 kg 13.15 kg 200 W 10 days Demonstrated High
PPT MEMS 8.80 kg 15.30 kg 100 W 30 days Research High
PPT Magnetic 8.95 kg 16.00 kg 100 W 30 days Demonstrated Low
Hall Colloid 8.95 kg 13.15 kg 200 W 10 days Research High
PPT μPPT 9.25 kg 15.50 kg 100 W 30 days Demonstrated High
N
2
H
4
N
2
H
4
9.15 kg 30.15 kg 5 W < 1 day Flight-Ready Medium
PPT Colloid 9.80 kg 15.55 kg 100 W 30 days Research High
MEMS MEMS 10.90 kg 50.45 kg 5 W < 1 day Research High
PPT PPT 11.50 kg 29.60 kg 100 W 30 days Demonstrated Medium
FEEP Magnetic 15.35 kg 16.10 kg 120 W 60 days Flight-Ready Low
FEEP MEMS 15.75 kg 17.45 kg 120 W 60 days Research High
FEEP μPPT 16.50 kg 17.70 kg 120 W 60 days Demonstrated High
FEEP Colloidal 16.60 kg 17.55 kg 120 W 60 days Research High

Table 3.5 – TechSat 21 Propulsion Options Comparison

3.4 Conclusions
If only the TechSat 21 flight experiment were to be considered, the combination of a conventional
hydrazine monopropellant thruster and micro pulsed plasma thrusters would be the clear choice.
The hydrazine system can deliver adequate for main propulsion within the specified mass budget
41
using off-the-shelf flight hardware, and the micro pulsed plasma thruster is the only demonstrated
microthruster capable of providing precision attitude control for near-term applications. This
combination minimizes both technical risk and system mass, which are the dominant concerns for
the first TechSat 21 flight.
However, chemical systems are completely unsuitable for follow-on operational missions, as the
higher ΔV requirement results in excessive propellant mass. For such missions, the preferred
option would be either a low-power Hall thruster or a pulsed plasma thruster for main propulsion
with a micro PPT system for attitude control. These systems meet all specified requirements save
mass, and exceed the mass budget by only a few kilograms. Furthermore, all of the technologies
involved have been demonstrated in larger-scale flight systems, with hardware appropriate for
TechSat 21 currently undergoing ground testing.
Since the Hall/μPPT and PPT/μPPT combinations can meet the requirements for near-term flight
experiments as well as future operational missions, it is likely preferable to design a single
propulsion system for high-performance microsatellites generally. This would minimize
development costs and allow early flight testing of the propulsive technologies needed for the later
mission. The chemical system could only be recommended for near-term flight experiments if
aggressive scheduling mandates an absolute minimum of technical risk to meet integration
deadlines.
This analysis indicates that, regardless of the choice of primary propulsion system, there is a
definite application for the μPPT as a microsatellite attitude control thruster. Full-scale PPTs
compete with Hall effect thrusters for this mission; as shown above the state of the art in low-
power HETs delivers much greater performance than comparable pulsed plasma thrusters. Thus,
42
the case for full-scale PPT development is less clear, and the μPPT should be developed as an
independent system rather than depending on the trigger circuit of a full-scale PPT.
While this analysis has focused on the TechSat 21 microsatellite, with implicit relevance to high
performance microsatellites in general, there are other potential applications for the PPT and
μPPT. For example, some mid-sized spacecraft have extremely precise stationkeeping
requirements due to their use of interferometric techniques at microwave, IR, and even visible
wavelengths
5
. Stationkeeping thrusters with a substantial and/or unpredictable minimum impulse
per firing, will introduce position disturbances comparable to the wavelength being observed,
rendering accurate interferometry impossible. Some interferometric spacecraft have adopted PPT
stationkeeping thrusters to minimize this effect.
Another, more speculative application, is the dynamic control of very large spacecraft. Some
future missions will require antennas, mirrors, or solar arrays of 100+ meters scale, but the mass
constraints of space launch will preclude the structural design techniques normally associated with
terrestrial structures of such scope. Large space structures are necessarily low in both stiffness and
damping, and minor disturbances can introduce large, persistent, possibly destructive oscillations.
Rather than attempt to prevent or mitigate such oscillations by structural design, it has been
proposed to actively damp them by means of microthrusters at appropriate locations on the
spacecraft structure. As with interferometric satellites, this application would require very small,
precise propulsive impulses and may be a mission for the PPT and/or μPPT.
However, we need not resort to such speculation. The TechSat 21 analysis indicates that there is a
class of missions for which the μPPT is the most suitable propulsion system presently available or
under development.
43

Chapter IV

PPT Performance Improvement

As determined in the previous chapter, the full-scale PPT in its present form is suitable only for
very specialized applications. For general use, it will not be competitive with other propulsion
systems unless its performance can be substantially increased. In particular, the electrical
efficiency must be at least doubled. This raises the obvious question – why is the efficiency so
disturbingly low in the first place?
4.1 Pulsed Operation – Loss Mechanisms
The PPT, unlike most advanced propulsion systems, necessarily operates in a pulsed mode. This
cannot help but reduce efficiency, in that transient start/stop losses are realized with every pulse
rather than merely once per period of thruster operation. Worse, as mentioned in chapter II, the
PPT undergoes a period of ringing following the initial discharge, resulting in further transient
losses with each current reversal within a pulse.
The obvious loss mechanisms here, are those associated with the creation of the magnetic field
surrounding the discharge electrodes. This magnetic field is necessary for the electromagnetic
acceleration of the propellant, but it represents a substantial investment of energy that will be used
only briefly. The magnitude of this potential loss is simply
E = ½ L I
2
(3.1)
For the LES 8/9 thruster, with a typical thruster head inductance of ~45 nH and a peak current of
18 kA, the stored energy of the magnetic field is 7.3 Joules. This represents 37% of the total
44
energy available per pulse, and is liable to be lost to ohmic heating and/or electromagnetic
radiation during the ringdown period.
Furthermore, the thruster head is not the only source of energy loss. While the PPT is constructed
to minimize parasitic inductance in the discharge path, there is a substantial internal inductance,
and resistance, in the capacitor. Also, the energy associated with the spark trigger is of no
propulsive value, and constitutes a pure loss with each discharge. Altogether, these losses amount
to roughly fifty percent of the discharge energy.
4.2 Pulsed Operation – Possible Solutions
If energy is being lost to repetitive magnetic field generation and collapse, three solutions appear.
First, the magnitude of the magnetic field energy can be reduced. Second, the magnetic field
energy can be recovered. Finally, the frequency with which the magnetic field is pulsed can be
reduced.
Reduction in magnetic field energy are constrained by the fact that a substantial magnetic field is
required to accelerate the propellant and produce thrust. However, only the magnetic field local to
the propellant face contributes to this mechanism. As can be seen in figure 4.1, the conventional
rectilinear geometry results in a magnetic field extending far from the propellant face. A coaxial
geometry can more effectively confine the magnetic field, reducing losses.

45
Figure 4.1 – Rectilinear vs. Coaxial Magnetic Field Geometry (Facing Thruster)
Figure 4.1 – Rectilinear vs. Coaxial Magnetic Field Geometry (Facing Thruster)
Recovering the magnetic field energy following a PPT discharge is difficult, but may be possible
to exploit the ringdown cycle for this purpose. The voltage at the first reverse peak, is typically
85-90% of the initial value. At this point, the magnetic field has collapsed and most of its energy
is returned to the discharge capacitor – albeit with reversed polarity. Normally, a series of
reversed discharge cycles will occur at this point, dissipating the remaining energy. If instead the
discharge circuit can be interrupted at this point, the energy will remain stored in the capacitor,
which would need only be “topped off” to recover full voltage for another, triggered, full-energy
discharge. As the voltage would be reversed, this would require a symmetric PPT circuit, which
should be readily achievable. It would also entail sacrificing that portion of the thrust generated
during the ringdown cycle, but as thrust generation during these lower-energy discharges is less
efficient than at full energy, this should be a net gain.
As interruption of highly inductive circuits is always difficult, another option is to simply cut the
capacitor out of the circuit. A crowbar diode can be installed across the capacitor, preventing it
from supporting a reverse voltage. Instead of ringing with discharges of alternating polarity, the
PPT will at the point of voltage reversal (and maximum current) switch from being an LRC-type

Electrodes
Propellant
Field Lines
Electrodes
Propellant
Field Lines
46
circuit to a simple LR circuit. The magnetic field and the discharge current will be sustained by
the inductance of the circuit, and an exponential decay will be observed.
The magnetic field energy will still be dissipated, but in a single continuous thrust-generating
discharge. Any extra inefficiency introduced by the periodic current reversals, can be avoided.
Figure 4.2 shows schematics of the classic, switched, and crowbarred PPT circuits, with figure 4.3
showing the expected current and voltage behavior during the discharge.
Figure 4.2 – Alternate PPT Circuit Schematics

Figure 4.3 – Alternate PPT Circuit Behavior
Crowbar Switched Standard
Voltage
Current

Cathode
Anode
Propellant
Capacitor
Standard PPT Switched PPT Crowbar PPT
Cathode
Anode
Propellant
Capacitor
Standard PPT Switched PPT Crowbar PPT
47
4.3 Propellant Utilization Inefficiencies
Another major source of inefficiency in conventional PPT operation is uneven propellant
acceleration. The energy efficiency of a propulsion system is highest when all of the propellant is
accelerated to the same velocity. In that case, thrust, specific impulse, and exhaust power are
given by
T = m
dot
V
e
(3.2)
I
sp
= V
e
/ g (3.3)

P
e
= ½ m
dot
V
e
2
(3.4)
If the propellant is instead accelerated in two populations A and B, each representing fraction f
A

and f
B
of the total mass flow, this becomes,
T = (f
A
m
dot
V
eA
+ f
B
m
dot
V
eB
) (3.5)
I
sp
= (f
A
V
eA
+ f
B
V
eB
) / g (3.6)

P
e
= ½ f
A
m
dot
V
eA
2

+ ½ f
B
m
dot
V
eB
2
(3.7)
Matching the performance of the single-component system requires that:
f
A
+ f
B
= 1 (3.8)
(f
A
m
dot
V
eA
+ f
B
m
dot
V
eB
) = m
dot
V
e
(3.9)
(f
A
V
eA
+ f
B
V
eB
) / g = V
e
/ g (3.10)
Combining equations 3.8 through 3.10 gives
48
f
B
= 1 – f
A
(3.11)
V
B
= V
e
+ (f
A
/(1-f
A
)) (V
e
-V
A
) (3.12)
And finally
P
e
= ½ m
dot
[ V
e
2
+ (f
A
/(1-f
A
)) (V
e
-V
A
)
2
] (3.13)
or P
e
= ½ m
dot
[ V
e
2
+ (f
B
/(1-f
B
)) (V
e
-V
B
)
2
] (3.14)
Since the (f
A
/(1-f
A
)) (V
e
-V
A
)
2
term is positive except in the trivial cases of V
e
= V
A
, f
A
=1, or f
B
=
1, any thruster whose exhaust consists of two finite streams of fluid at different velocity, will
require more exhaust power than a system delivering the same thrust and specific impulse with a
single, constant-velocity exhaust stream. In the limiting case where exhaust stream A is of zero or
negligible velocity, we get
P
e
= ½ m
dot
V
e
2
[1 + (f
A
/(1-f
A
))] (3.15)
or P
e2
= [1 + (f
A
/(1-f
A
))] P
e1
(3.16)
This is of particular relevance to the PPT, because it is estimated that 80-90% of the exhaust
stream is of zero or negligible velocity
17
. Emission spectroscopy in the exhaust plume shows
substantial emission of neutral vapor for a period of several hundred microseconds following the
discharge. This neutral vapor cannot be electromagnetically accelerated, not only because of its
own neutrality but because the discharge current and magnetic field strength have long since
decayed to negligible values. It can undergo thermal expansion and thus develop a velocity
component perpendicular to the propellant face, but this will result in a velocity of no more than
49
1000 m/s, compared to the ~10,000 m/s implied by the PPT’s specific impulse of 1000 s. This
neutral vapor emission is estimated at 40% of the total propellant consumed per pulse.
Worse, a significant fraction of the propellant is ejected in the form of solid particles at even
lower velocities
17
. These particles can be seen in high-speed camera images of the thruster in the
period following the discharge, and comparing photographic streak length with exposure time
suggests a velocity of 2-300 m/s.
Determining the total propellant mass associated with this particulate emission, required collection
and analysis of the particles. For this purpose, witness plates consisting of 4” diameter
semiconductor-grade silicon wafers was placed downstream of an XPPT-1 thruster, at distances
varying from 10-15 cm and angles off-axis of zero to ninety degrees. The thruster was fired for
1000 pulses, and selected areas of the witness plates were examined under optical and electron
microscopes. The number and size of particles was counted, and the elemental composition
determined by X-ray backscatter from the electron microscope.
The particle emission was determined to consist of two populations. One, comprised mostly small
(~1 micron), spherical particles of predominately iron composition. As the particular thruster used
in this experiment used stainless steel electrodes, this probably represents material ablated from
the electrode surface during arc formation. The more significant population, however, consists of
irregular particles of ~100 micron scale and composed primarily of carbon and fluorine. This is
clearly propellant, and the particle count and size integrated over the collection area suggests that
these particles represent ~40% of the total propellant consumption
21
.
The mechanism for the neutral vapor and particulate emission is believed to be associated with
deep heating of the propellant during the discharge. Teflon is optically translucent in the visible
50
and near-UV regimes, and composed of low-Z materials that are ineffective in blocking soft X-ray
radiation. The bulk of the energy radiated by the arc discharge will therefore penetrate to a
substantial depth in the propellant face, on the order of 200 microns
1,18
.
By comparison, only the outermost ~100 nm of propellant is actually ablated during each pulse,
and only 10-20% of that is ionized and electromagnetically accelerated in the discharge period. If
energy is being deposited much deeper than this in the propellant body, and if the propellant face
is being sufficiently heated to drive direct, rapid ablation during the discharge, it seems likely that
there is a sufficient reservoir of propellant near the newly-exposed face and heated to temperatures
close to the boiling point, that significant evolution of vapor will occur in the low-pressure
environment of the post-discharge period. Furthermore, any imperfection in the propellant can
serve as a nucleation site for internal boiling, resulting in vapor microexplosions that would eject
substantial chunks of solid or non-boiling liquid propellant. The approximate correspondence
between the deep heating scale length and the ejected particle scale length suggests this is the
source of the observed particulate emissions.
4.4 Propellant Inefficiency – Poential Solutions
If indeed deep heating is responsible for the neutral vapor and particulate inefficiencies, it should
be possible to suppress this loss mechanism by using a propellant with greater opacity at the
relevant wavelengths. There are sound reasons for using Teflon™ as the propellant in a PPT, but
a small concentration of dopants such as graphite (for visible or near-UV) or molybdenum
disulfide (far UV and soft X-ray) should greatly increase opacity and suppress deep heating.
Furthermore, Teflon may not in fact be the best polymer for use as the bulk of the propellant.
51
Materials with lower vapor pressures and/or greater mechanical strength may be less susceptible
to sublimation at the surface or nucleated boiling in the interior.
The ultimate extreme would be to use a dense, metallic propellant. This has the obvious problem
that metals, being conductive, will tend to short the PPT electrodes circuit and prevent the
development of an arc discharge – or even the charging of the PPT capacitor. However, as one of
the remedies proposed in section 4.2 for magnetic-field loss issues is a PPT with an actively
switched discharge current, this problem may not be insurmountable.
An alternative to suppressing deep heating and neutral propellant vapor emission is to exploit it for
thrust. Free expansion of the propellant vapor will result in velocities of only a few hundred
meters per second, but expansion through a converging/diverging nozzle may offer better
performance – this is, after all, the primary thrust mechanism of a chemical rocket. If the
particulate fraction can be entrained in this flow, or better still vaporized along the way, it too will
contribute.
This requires that the PPT exhaust be channeled through a rocket nozzle. Again, in section 4.2 we
are already contemplating a coaxial geometry; with minor modifications the outer electrode can
serve as a converging-diverging nozzle. Also, propellant chemistry can be modified to support
this operating mode. Best performance would be achieved with a propellant of high evaporation
or sublimation temperature and low molecular weight. Ideally, the propellant would have a
positive heat of formation to provide extra thermal energy and to assure breakdown of solid
particles and long polymer chains to low-MW vapor.
Propellants considered for advanced PPT testing include Teflon™ and Teflon analogs, POSS (a
proprietary polymer developed at AFRL and expected to have a low vapor pressure), ASPEN-109
52
(a proprietary composition with positive heat of formation), Teflon and POSS with up to 5%
graphite powder, Teflon and POSS with up to 5% molybdenum disulfide powder, and metallic
bismuth. The latter is unique among metals in its relatively low electrical conductivity, making it
the most suitable for the admittedly speculative prospect of a metallic-propellant switched PPT.
As a parallel effort by Burton at the University of Illinois
6,8
focused on developing a PPT to
thermally accelerate neutral propellant vapor, that potential solution was not pursued in this effort.
4.5 The XPPT-48 Testbed
In order to test these potential solutions, a new pulsed plasma thruster was developed and tested.
The key characteristics of this thruster are, coaxial geometry, minimum internal inductance and
resistance, provision for an SCR switch and/or crowbar diode, and provision for use of alternate
conductive and non-conductive propellants. Minimum cost and development effort, through the
use of off-the-shelf components, was of course also a consideration. The thruster was designed to
operate efficiently at pulse energies of 20 to 60 joules; the former to facilitate comparison of test
results with previous PPT experience, the latter to allow scaling to higher-power operation as
seems likely to be required in future applications.
The critical parameter for PPT scaling, is believed to be discharge energy per unit area over the
propellant face. A certain minimum energy density will be required to effectively drive propellant
ablation. However, excessive energy density would probably aggravate the late-time vaporization
problem. Thrusters such as the LES 8/9 and XPPT-1, with one-inch propellant faces and
discharge energies of 20-25 Joules, have areal energy densities of 31-38 millijJoules per square
millimeter. Matching this for a coaxial geometry with a nominal discharge energy of 40 Joules
and a 2:1 ratio of electrode diameters, calls for an outer electrode diameter of 48 mm. Other
53
parameters of the XPPT-48 design (and incidentally the thruster’s name) were determined from
this key dimension.
The ability to incorporate power electronics for switched and crowbar operation is provided by
incorporating space for hockey-puck power transistors, SCRs or diodes in series or parallel with
the discharge path. When not required for power electrodes, these spaces can be left empty or
filled with copper spacer rings. Provision is also made to mount a Rogowski coil for current
diagnostics. Conductive-propellant operation is supported by using a switch SCR or power
transistor in series with the discharge, triggered simultaneously with the PPT discharge. PPT
triggering, with conductive or non-conductive propellants, is accomplished by a coaxially
positioned Unison helicopter turbine. The thruster can also use any of a wide range of oil-filled
capacitors to allow variation of discharge voltage and energy.
Figure 4.4 is a cross-section of the XPPT-48 design, with figure 4.5 a photograph of the asssmbled
thruster with a Powerex C712L SCR and Rogowski coil. Figure 4.6 is an electrical schematic of
the XPPT-48 showing optional SCR and crowbar diode locations.

Figure 4.4 – XPPT-48 Cross Section
SCR
Propellant
Capacitor
Diode
Current
Sensor
54

Figure 4.5 – Assembled XPPT-48 thruster

Figure 4.6 – XPPT-48 thruster schematic

+HV
Optional SCR
Optional
Crowbar Diode
Optional Current
Monitor
Discharge
Capacitor
Trigger Circuit
Spark Plug
Propellant
Propellant
55
Functional testing of the XPPT-48 demonstrated reliable performance in normal operation. A
comparison of the performance of the XPPT-48 with that of the LES 8/9 PPT is given in table 4.1.
As can be seen, there are modest performance gains, presumably due to the coaxial geometry.
These gains are well below the substantial increases required to justify continued PPT
development; for that, semiconductor switching and/or alternate propellants will be required.
LES 8/9 PPT XPPT-48, 20J XPPT-48, 40J XPPT-48, 60J
Thrust, mN 0.30 0.32 0.63 1.00
Isp, s 1000 1435 980 1125
Power, W 25.5 25 50 75
Efficiency, % 5.75% 9.0% 6.05% 7.35%
Min I
bit
, μN-s 237 320 630 1000
Dry Mass, kg 6.6 10.5 10.5 10.5
Life, hours 9500 ? ? ?
Table 4.1 – XPPT-48 performance
Unfortunately, attempts to operate the XPPT-48 in a switched mode, consistently resulted in
failure of the semiconductor switch. Even heavy-duty Powerex C712L SCRs were not able
to survive the voltage transients associated with switch opening at first current reversal.
Thruster lifetime in this operating mode was limited to a few dozen pulses, which was not
sufficient to measure performance.
4.6 Alternate Propellant Operation
Laboratory research has shown that the traditional PPT propellant, Teflon, is a strong
contributor to the poor performance (~7% efficiency) of the device
20
. For example,
particulate expulsion (attributed to the Teflon fabrication process) and late–time vaporization
(due to either deep energy deposition from x–ray transmission enabled by the low–Z atoms
in the Teflon polymer chain, or by simple heat transfer) combine to waste up to 90% of the
56
PPT propellant while creating essentially zero thrust. In addition the plasma formation
energy for Teflon is much greater than that of other electric propulsion propellants – largely
due to heat of vaporization, depolymerization, and the high ionization energy of C and F.
Xenon, the most commonly used fuel for Hall thrusters and ion engines, requires 834 Joules
per gram to ionize. Ionization to Xe
+2
and Xe
+3
requires only 2.5 and 4 times as much energy
respectively. Teflon, on the other hand, requires 1.5x10
5
Joules per gram
18
. It has long been
recognized in PPT research and development that replacing the Teflon propellant could be
an attractive path towards enhanced performance. However, experiments over the last 30
years have shown every alternate propellant to fail by charring, which leaves a carbon
conduction path for the arc; melting, which changes the length of the conduction path;
mechanical failure in the fuel feed machinery; or some other means.

Only Teflon has demonstrated the electrical and mechanical robustness needed to survive the
harsh environment of the PPT discharge; hence Teflon has been used in every PPT flight to
date. Based on the previous work it became obvious that research dedicated towards
electronics and circuitry of the PPT rather than the physical layout would be a better use of
limited resources.

A class of alternative fuels for PPTs consists of controlled chemically energetic solid
propellants, or controlled solids. These are proprietary solid rocket propellants, with stored
energy of approximately 2.5 MJ/kg, which will combust only in the presence of an electric
current. Chemically energetic propellants are attractive for PPT application for two reasons.
First, the additional chemical energy may add to the performance of the thruster through
gasdynamic acceleration. This is likely to be particularly important during the late-time
57
vaporization phase, when a large quantity of propellant is accelerated only through
gasdynamic effects. Second, the positive heat of formation of such a propellant will promote
complete decomposition under circumstances where other alternate propellants would
merely char.

The XPPT-48 incorporates a switched PPT circuit designed to circumvent the primary
failure mode of alternate propellants: charring at the propellant face leading to a conductive
short circuit. Switched PPTs also have the benefit of enabling the use of conductive
propellants. In a standard PPT, the on–board capacitor is charged to about 1500 V and 20 J.
This voltage appears at the electrodes and is held off by the Teflon™ fuel/insulator until a
separate trigger circuit is fired to discharge a low–energy (.5 J) spark plug embedded in the
PPT cathode. In a standard PPT, even a mildly conductive propellant would short the
capacitor and prevent it from charging.

In a switched PPT, the capacitor is separated from the electrodes by a switch. This could be
a mechanical switch, but in a space application there should not be any moving parts so an
electronic switch is preferred. No charge appears across the propellant face during the
charging of the capacitor, but is instead applied in a burst fashion when the switch is closed.
A short–circuit at the propellant face will not stop the discharge, and over time the transient
nature of the applied current is likely to result in the conductive carbon being vaporized
away.

For the chemically augmented PPT, the XPPT-48 was used to test the electrically-conductive
controlled solids. Non–conductive formulations were also tested using conventional PPTs.
58
The use of conductive propellants was expected to increase control over the mass ablation
rate by judicious choice of pulse energy and transient skin depth at the propellant face.

The applicability of chemically energetic propellants to the PPT, however, depends on the
ability to exploit gasdynamic acceleration during the late-time vaporization phase. The
XPPT-48 has only a limited capability in this regard, and other thrusters
6,8
designed
specifically to exploit gasdynamic effects have not been strong performers. This technology,
while promising in other areas, does not seem to be the solution to the PPT efficiency
problem.

Furthermore, the propellant formulations available for this work, were only marginally suitable for
PPT operation. Additional formulations were developed as the fuel was examined and found
wanting in various respects. Certain failings in the fuel included a low melting point, which
would result in melting of the propellant during tests of even short durations. Additionally, the
electrical properties of the fuel were often nonrepeatable and even inconsistent within a single
batch, resulting in uneven propellant expenditure. These problems were never adequately
overcome.
4.7 Effect of Variable Capacitance and Discharge Energy
One of the attractive features of the XPPT-48 is the speed and ease of replacing electronic
components, particularly the capacitor. So it is ideally suited to investigate how the change
in capacitance affected the ablation rate at different energies. If a dramatic change was
discovered, future work could ascertain whether or not a comparable change in thrust was
present.
59

The two NWL oil–filled capacitors chosen measured 19.7 and 42.0 µF. Capacitance was
verified using a Sencore LC–102 capacitance analyzer that was calibrated to a stated
accuracy of 0.05%.

While a few test runs were done at other power levels, most measurements were performed
at 40 Joules. Each test consisted of at least 10,000 shots at the target power level. All tests
were done at a repetition rate of 1Hz. Because of this requirement, power levels over 60 J
were beyond the ability of our power supplies to recharge the capacitor in time to initiate the
next pulse. The thruster was fired for several thousand shots to ensure a statistically
significant ablation rate measurement. Table 4.2 details the test results in the order obtained.

19.7 μF 42.0 μF
20 J 22.5,23.1 μg/pulse 24.1 μg/pulse
40 J 63.1, 62.6, 69.1, 74.1, 72.4, 72.5 69.2, 67.6, 57.2, 54.6, 57.7
60 J 86.4,87.9 97.4
Table 4.2: Ablation Rates as a Function of Power Level [μg/pulse]

With each increase in discharge energy there is a substantial change in ablation, which is to
be expected. However, the range of measured ablations at each energy remained fairly
stable, suggesting that varying the capacitance will have little to no effect on the thruster
performance if the energy remains constant.

Additionally, tests measuring the discharge current with a Rogowski coil were performed at
both capacitances at energies from 10–90 J. The data presented in figures 4.7 and 4.8
indicate that, unsurprisingly, lower capacitance (and, consequently, lower inductance) at like
energies corresponds to higher peak current and faster current dissipation.
60
Figure 4.7: XPPT-48 Current Traces, 19.7 μF
Figure 4.8: XPPT-48 Current Traces, 42.0 μF

-40
-30
-20
-10
0
10
20
30
0 5 10 15 20 25 30 35 40
Time [? s]
Current [kA]
10 J
20 J
30 J
40 J
50 J
60 J
70 J
80 J
90 J
-40
-30
-20
-10
0
10
20
30
0 5 10 15 20 25 30 35 40
Time [? s]
Current [kA]
10 J
20 J
30 J
40 J
50 J
60 J
70 J
80 J
90 J
61
Since higher energies can be achieved with greater capacitance without exceeding our
voltage–supply capabilities, further traces were done at 42.0 μF in the 100–200 Joule range.
These are presented in Figure .9. Unfortunately, thruster heating at these increased
discharge energies caused bulk propellant melting before performance could be measured.

Figure 4.9: Higher Energy XPPT-48 Current Traces, 42.0 μF

4.8 Conclusions
Put simply, none of it works. In most cases, the proposed “solutions” not only failed to increase
thruster performance, but resulted in a thruster that failed to operate at all. The XPPT-48 did, as
mentioned, operate successfully in the standard configuration with Teflon™ propellant, but
offered only slightly greater performance (9.0% efficiency vs 5.75%) than the LES 8/9. This gain
-50
-40
-30
-20
-10
0
10
20
30
0 5 10 15 20 25 30 35 40
Time [? s]
Current [kA]
100 J
125 J
150 J
175 J
200 J
62
perhaps validates the choice of a coaxial geometry, but does not offer the revolutionary increase in
efficiency necessary to make the full-scale PPT a competitive thruster.
Operation of the XPPT-48 in switched mode, proved remarkably effective at destroying expensive
power electronics. It also proved expensive at destroying expensive diagnostic systems being
used to determine why the expensive power electronics were being destroyed. Insofar as this was
an effort to deliberately open a high-current inductive circuit, even at a momentary zero-current
transient, this result is not terribly surprising. It was sufficient motive for us to stop trying that
solution even before we had learned enough to offer a clever explanation of why we ought to stop
trying,
The Burton effort at the University of Illinois
18
tested the crowbar operating mode in two different
PPT designs, one optimized for thermal acceleration of neutral propellant vapor. Both thrusters
were at least functional, but delivered performance inferior to the LES 8/9.
The XPPT-48 was tested with all of the alternative propellants mentioned in section 4.4, above.
With the exception of ASPEN-109 and Bismuth, all of these propellants demonstrated extreme
charring of the propellant surface after only a few thousand discharges. The electrically
conductive char served to short the PPT electrodes, preventing operation of the device. Absent a
reliably-switched PPT circuit, electrically conductive propellants cannot be used. This applies
even to the relatively low conductivity of metallic bismuth; it was extremely difficult to achieve
any sort of discharge using bismuth propellant in the unswitched XPPT-48. ASPEN-109
demonstrated reliable operation with no charring even over long periods, but offered no increase
in performance.
63
Finally, consider figure 4.10, below. This shows the performance of every PPT for which
sufficient data was available. The thrusters included cover nearly three orders of magnitude in
scale, and include nearly every geometry and operating mode considered for the PPT.

Figure 4.10 – PPT Performance Trends
All of these thrusters can be characterized by a nearly linear relationship between the thrust:power
ratio and the log of the discharge energy. To the extent that performance increases are possible
for the PPT, they will come only by greatly increasing the discharge energy. This would put the
PPT in the high-power thruster regime, where it would compete directly with full-scale Hall
thrusters, arcjets, ion thrusters, and the like, which are clearly superior in that regime.
We, and others, have made our best effort to increase the performance and efficiency of the PPT.
This does not seem to be possible with any reasonable effort or foreseeable near-term advances.
The PPT is a thruster for niche applications where efficient conversion of electric power to thrust
is not a requirement.
Thrust/Power
Discharge Energy (J)
0
10
20
30
40
1 10 100 1000
LES 6
AF Millipound
LES 8/9
EO-1
Busek, Single-Shot, ¼” µPPT
5 µN/Watt – Min thrust promised to TechSat21, FalconSat 3
AFRL, ¼” µPPT
MicroPPT
Other PPTs
Benchmark PPTs
Thrust/Power
Discharge Energy (J)
0
10
20
30
40
1 10 100 1000
LES 6
AF Millipound
LES 8/9
EO-1
Busek, Single-Shot, ¼” µPPT
5 µN/Watt – Min thrust promised to TechSat21, FalconSat 3
AFRL, ¼” µPPT
MicroPPT
Other PPTs
Benchmark PPTs
64

Chapter V

The Two-Stage μPPT

If full-scale PPTs are unlikely to be competitive, applications may remain for the μPPT.
However, much of the perceived merit of the μPPT has traditionally lain in the synergy between
PPT and μPPT in a combined system, using a common trigger circuit. We must now consider a
spacecraft operating only μPPT thrusters, or μPPTs in conjunction with some wholly different
main propulsion system. In this context, the use of a switched trigger circuit of any sort is now an
undesired complexity. Furthermore, it is no longer possible to assume that (relatively) high-
impulse propulsive or stationkeeping maneuvers will be handled by another propulsion system.
The μPPT may be the only propulsion system on the spacecraft, and thus total impulse
requirements may increase substantially.
Unfortunately, these new constraints are mutually contradictory. Eliminating the trigger circuit
would require operating in a self-breakdown mode, which if off-the-shelf electronics are to be
used will probably be limited to 1-2 kV. Increased total input, requires an increased propellant
load, which will require either increasing the diameter of the propellant/electrode assemblies or
switching between a large number of small-diameter assemblies. Increased propellant/electrode
diameter in a self-breakdown mode calls for greatly increased discharge voltage, on the order of
tens of kilovolts for a ¼” diameter. Actively switching between multiple propellant/electrode
assemblies, would require complex power electronics.

65
5.1 Two-Stage µPPT Concept
After considering many potential solutions to this dilemma, the two-stage μPPT was
developed. This improvement on the standard µPPT , pictured in figure 5.1. uses a three–
electrode circuit as shown schematically in figure 5.2
15
. A small–diameter conductive rod is
used as a “center electrode” and is surrounded by a small diameter annulus of Teflon™
propellant. This assembly is then encased in a conductive tube that acts as an “intermediate
electrode.” The intermediate electrode is surrounded by a second, larger, annulus of
Teflon™, which is encased in a larger diameter outer conductive rod that acts as the “outer
electrode.”

Figure 5.1: Two-Stage µPPT

66

Figure 5.2: Two-Stage µPPT Schematic

The µPPT is fired by a low–energy breakdown between the intermediate and central
electrodes. This discharge provides enough seed ionization to enable the higher energy
conduction breakdown between the intermediate and outer electrode. The discharge
between the intermediate and central electrode is referred to as the “trigger discharge.” The
discharge between the intermediate and outer electrode is referred to as the “main
discharge.” Typically the trigger discharge energy is about 1/50th that of the main
discharge.

When the PPU is commanded on, the µPPT circuit is designed to energize the capacitors as
shown in Figure 5.3. The main capacitor quickly charges to the full output voltage of the
PPU. The charge voltage on the main capacitor is designed to be well below the threshold
for surface discharge initiation between the intermediate and outer electrode so discharge
C
trigger
C
main
(98% of Energy)
R
trigger
Center
Electrode
Intermediate
Electrode
Outer
Electrode
Teflon Annulus
Spacecraft
Ground
PPU
67
initiation is precluded. The charging rate of the trigger capacitor is greatly slowed by the
large resistance in the charge path. Nominally, when the trigger capacitor reaches its
threshold voltage for surface breakdown a discharge arc between the center and intermediate
electrode will initiate. This small low–energy discharge will provide sufficient seed
ionization near the propellant face that the main discharge will be initiated.

Figure 5.3: Energizing the Two-Stage µPPT Capacitors
Using the three–electrode design has three major benefits. First, the energy of the main discharge
now has minimal shot–to–shot variation, decreasing the likelihood of carbonization on the
propellant face. This should also have the effect of the µPPT delivering a more uniform impulse–
bit if it were to be used in a mission requiring single shot operation. A damped–sinusoid thrust
stand would be required to quantify statistical variation in the single–shot impulses. Second, the
seed ionization from the trigger discharge greatly reduces the voltage required on the main
discharge. For example, for a 0.25” diameter outer electrode in a 2–electrode configuration, up to
40 kV is required to reliably initiate the discharge. With the three–electrode design this voltage is
typically 3 kV or less. This significant reduction in voltage greatly reduces the design
Charge Voltage
Time
Main Capacitor
Trigger
Capacitor
(τ ~RC
trig
)
Trigger Threshold Voltage
Nominal Firing T ime
Charge Voltage
Time
Main Capacitor
Trigger
Capacitor
(τ ~RC
trig
)
Trigger Threshold Voltage
Nominal Firing T ime
68
requirements for the PPU. Third, the design is far more robust to short–term increases in the
voltage required to initiate the trigger discharge.
5.2 Two-Stage µPPT Fabrication
Several dozen two-stage μPPT systems were fabricated over the course of this effort, both to
testvarious design modifications and to demonstrate repeatability. This was facilitated by
designing the baseline μPPT for assembly from a small number of off-the-shelf components.
Eventually, it was possible to assemble a new μPPT in a matter of hours. The baseline
configuration, upon which all testing in this section is based, is detailed in Table 5.1.

Initially, the thrusters used were made from Pasternack RG–401 and PE–047SR semi–rigid
cable. The 0.250”–diameter RG–401 had its center conductor drilled out and the 0.047”–
diameter PE–047SR was inserted, forming a three–electrode system. While this sufficed for
early lab tests, the construction process was too time consuming, labor–intensive, and prone
to failure.
Table 5.1: Baseline Two-Stage µPPT Configuration
Outer Electrode Diameter 0.141”
Outer Electrode Material Copper
Intermediate Electrode Diameter 0.034”
Intermediate Electrode Material Copper
Center Electrode Diameter 0.008”
Center Electrode Material Silver–Covered, Copper–Clad Steel
Main Capacitance 0.5 µF
Trigger Capacitance 0.02 µF
Trigger Resistance 100 M
Power Supply EMCO E40 4kV
69

To alleviate this, Precision Tube Company of Salisbury, MD was contracted to produce a
custom semi–rigid cable, designated TRX 141–034–50–50, to the specifications detailed in
Table . This coaxial cable became essentially a mass–produced µPPT tube. An end–on
view of the cable is presented in Figure 5.4. Some use was still made of the hand-fabricated,
0.250”-diameter design in order to explore the effect of propellant diameter.

Figure 5.4: µPPT Tube Custom–Made by Precision Tube Company

Similarly, off–the–shelf mica paper capacitors were initially used during basic testing.
These 0.1 and 0.01 µF capacitors were part of the Custom Electronics, Inc. standard line of
space–rated products and rated at 8 and 10 kilovolts, respectively. Once it had been
established that 0.5 and 0.02 µF capacitances were ideal for the µPPT, we requested they
combine the two capacitors into a single unit with a common ground. Additionally, to save
mass, they were rated to the peak energy the µPPTs would be fired at: 3 kV, corresponding
70
to 2.25 Joules. The resulting yellow capacitor, along with the original white Custom models,
is shown in figure 5.5

Figure 5.5: Custom Electronics, Inc. Capacitors

The three leads on the yellow capacitor correspond to, going clockwise from the upper left:
1: Main Capacitance (0.5 µF), C: Common (ground), and 2: Trigger Capacitance (0.02 µF).

The remaining two electrical components were truly off–the–shelf. The 100 MΩ resistors
were Ohmite Mfg. Co. Precision Thick Film Axial Leaded High Voltage/High Resistance
Resistors, part of their standard Maxi–Mox line. The EMCO High Voltage Corporation E40
71
power supply takes in a 0–15 V input and scales the output from 0–4 kV, pushing a
maximum current of 0.75 mA. The EMCO supply is pictured in figure 5.6.

Figure 5.6: EMCO E40 4kV Supply
For laboratory test purposes, the thruster is assembled by simply hand-soldering the components
together, using copper stripline for the main discharge path and 24-gauge aluminum wire
elsewhere. Some thrusters also incorporated packaging, internal reinforcement, and external
interfaces approximating those of a flight system, to support realistic testing. For either version,
operation of the thruster requires only the application of a ~12 VDC current to the EMCO power
supply, whereupon the thruster will discharge at a ~1 Hz rate. A completed μPPT photographed
in mid-discharge is shown in figure 5.7
72

Figure 5.7: Baseline Two-Stage µPPT (without external packaging)

5.3 Test Facilities
Existing test facilities at AFRL were inadequate for μPPT development in two respects. First,
scheduling and cycle-time constraints on the large vacuum chambers required experiments to be
planned weeks in advance. This was a critical bottleneck in the development of a thruster that
could be designed and built in a matter of days. Traditional thruster development has been
supported by extensive analytical and numerical simulation efforts with only limited
experimentation. The physics of the μPPT are quite complex, involving phase and density
changes all the way from the solid state to low-density plasma and neutral vapor, and defy ready
analysis or simulation. Thus, it is necessary to exploit the small scale and simplicity of the μPPT
for a rapid design-build-test cycle.
73

Second, the most precise thrust stand at AFRL, a NASA-Glenn/Haag torsion pendulum design,
had a precision under ideal circumstances of +/- 200 μN. This is comparable to the maximum
thrust of a μPPT, rendering useful thrust measurement nearly impossible. Attempts to improve
the performance of the NASA/Lewis thrust stand had in the past met apparently insurmountable
obstacles due to differential thermal expansion of the thrust stand components, and incomplete
isolation of external vibration noise from the thrust measurement.

To support rapid functional testing of the μPPT, two new vacuum chambers were constructed by
WER, both physically located at the AFRL facility on Edwards Air Force Base. In line with
AFRL’s established nomenclature, these are identified as Chamber 8 and Chamber 9. These
chambers are shown in figures 5.8 and 5.9, respectively.

Figure 5.8: Chamber 8 (Glass Bell Jar)
74

Figure 5.9: Chamber 9 (Steel Bell Jar)
Chamber 8 consists of an 18” x 30” Pyrex bell jar with a Varian 300–SF turbomolecular pump
backed by a Varian SD–301 roughing pump. It has eight access ports in the jar base, which add
the capability of testing up to seven thrusters simultaneously. The chamber can achieve an
ultimate pressure of 0.5 µTorr, which betters that of existing chambers 2 and 5A. Additionally,
the time to full vacuum is well under an hour, providing for multiple testing sessions daily.

Chamber 9, was built from a 24” x 30” stainless steel bell jar with a Varian 300–SF
turbomolecular pump backed by a Varian SD–301 roughing pump and is built in the same
configuration as Chamber 8.
75

For thermal-vacuum testing, the existing Chamber 5A was modified by the addition of a thermal
shroud. The thermal shroud, built by Imco, uses a heat exchange system fuel by liquid nitrogen to
achieve a temperature range from –150°C to +180°C. A thermally–controlled mounting plate
inside the shroud simulates the spacecraft panel on which a PPT would be mounted, effectively
mimicking the heat sinking capabilities of an on–orbit satellite. Using ethylene glycol, the plate
temperature ranges from –30°C to +100°C. The shroud and plate are shown in figure 5.10.

Figure 5.10: Thermal Shroud in AFRL Chamber 5A
The requirement for thrust measurements of microNewton precision, was met by a fundamental
change in the operating mode of the NASA-Glenn/Haag thrust stand in AFRL chamber 2.
Chamber 2, which is 2.4–meters in diameter by 3.8–meters long and backed by 2 diffusion
pumps, is shown in figure 5.11. This chamber unfortunately requires over five hours to pump
down, leading to a longer test cycle when thrust measurements are required.
76
Figure 5.11: Chamber 2 (with Thrust Stand)

5.4 The MicroNewton Thrust Stand
Effective thrust measurement at microNewton levels was accomplished by modifying the
NASA-Glenn/Haag thrust stand to operate in a forced resonant oscillation mode. Normally,
the thrust stand operates by having thruster operation displace the indicator from a known,
stationary reference position. An electromagnetic damper is used to maintain a constant
deflection in spite of unsteady thruster operation, start/stop transients, and external
vibrations. For microNewton thrust measurements, the thrust stand is operated in an
oscillating, undamped or weakly damped mode. The thruster is then fired in resonance with
the swing arm oscillation, such that the thruster is engaged for half of each thrust stand
77
period. A schematic of the AFRL Forced–Harmonic Oscillator Thrust Stand is presented in
figure 5.12.

This operating mode is available only to thrusters that operate in a pulsed mode, or can be
started and stopped precisely, with negligible transient effects, on a timescale corresponding
to the ~4-second half-period of swing arm oscillation. Calibration of the thrust stand
similarly requires periodic, rather than steady state, application of a calibration force. Where
in normal operation the thrust stand uses fixed weights hung from a pulley to provide a
steady calibration force, in harmonic oscillation mode an electromagnet is used to
periodically lift and release small magnetic calibration masses of known weight.

Figure 5.12: AFRL Forced–Harmonic Oscillator Thrust Stand Schematic

Because the thrust is resonantly applied with swing arm motion, the amplitude of oscillation
is amplified, facilitating measurement. The motion of the swing arm behaves as a forced
78
harmonic oscillator, with appropriately insensitive response to external vibrational forces
that are either random in nature, such as people walking in the vicinity of the operating thrust
stand, or are periodic but are of sufficiently different frequency that energy does not
efficiently couple to the motion of the swing arm, such as operation of vacuum pumps.

Further, this system is insensitive to thermal drift as long as the total swing arm motion
remains within the linear range of position detection. This is due to the relative measurement
of noting the difference between oscillation maxima and minima to produce swing arm total
amplitude, which does not vary if the center of oscillation drifts in time.

Figure 5.12 shows how the thrust stand increases amplitude to a steady value with an applied
force every half–period.

time
LVDT swing arm position
drift from
zero
maximum
minimum
amplitude
-5 V 0 V + 5 V
time
amplitude
0 V 10 V
b)
a)
time
LVDT swing arm position
drift from
zero
maximum
minimum
amplitude
-5 V 0 V + 5 V
time
amplitude
0 V 10 V
b)
a)

Figure 5.13: Thrust Stand Oscillation, Amplitude, and Long Term Drift
79

A sample calibration is shown in figure 5.14. Periodic application of calibration weights of
varying masses increase the amplitude of the thrust stand to converged values. It should be
noted that this level of performance required the incorporation of a weak but deliberate
damping force to the system. A truly undamped oscillator, of course, would never converge,
and using only the intrinsic damping of the torsion-pendulum thrust stand resulted in erratic
convergence. Thus, a dashpot-type damper consisting of a moving stylus suspended in a
fixed container of vacuum oil was incorporated in the system.

Figure 5.14: Calibration Behavior

Figure 5.15 shows the linearity of the calibration, which is directly related to the accuracy
(and consequently uncertainty) of the thrust stand.
0
1
2
3
0 1000 2000 3000
time (seconds)
Amplitude (V)
1st mass
2nd mass
3rd mass
4th mass
80

Figure 5.15: Resulting Linear Calibration

Obviously, operation of the thrust stand in harmonic oscillation mode cannot be performed
manually. The periodic application of thrust or calibration force must be controlled to a fraction of
a second, every eight seconds, for a period of up to an hour. After a few proof-of-concept
experiments, it was necessary to automate this process. The bulk of the work of automating the
thrust stand was done by technical staff at AFRL, with particular credit to Dr. Mike Dulligan and
Capt. James Lake.
The automation platform was a Dell® Dimension 8200 desktop PC, with an Intel® Pentium
IV processor operating at a speed of 2500 megahertz. Relevant software included the
Windows XP Professional® operating , the National Instruments LabVIEW 6.1®
programming environment, and various coommunication drivers for acquisition and control
hardware are provided by the National Instruments Measurement and Automation Explorer
2.2®. Peripheral hardware necessary to control the apparatus consisted of a National
y = 19.263x - 0.0531
R
2
= 0.9997
0
25
50
75
01 23 4
amplitude (V)
force (uN)
81
Instrument MID–7604 stepper motor driver, a PXI–7344 Motion Controller, a PXI–2565
Relay Switch, and a National Instruments BNC–2090 digital voltage reader.

Several discrete automation drivers were written and incorporated into a single LabVIEW®
Virtual Instrument controlling all aspects of thrust stand operation from initial calibration to
thrust measurement and, if necessary, troubleshooting and diagnostics. Once the system was
validated, true pushbutton operation was achieved and operator supervision was not required
during calibration or thrust measurement.

Initial validation of the thrust stand was performed with an externally–triggered, 6.35 mm
diameter, two–electrode µPPT. The 2.94 µF Maxwell capacitor was fully charged at all
times, and the thruster breakdown was initiated by a 1.59–mm, two–stage, trigger µPPT
operating solely as a seed plasma source. The trigger µPPT was placed facing the µPPT,
approximately 25 mm from the thruster exit plane, and it was fired at 5 kV at a stored energy
of 0.45 J at a frequency of 1 Hz. This was done to strictly control the half–period–on, half–
period–off operation of the thrust stand without having to be concerned about capacitor
charging rates and extraneous discharges during the reverse motion of the stand.

These validation tests were performed at pulse energies of 3-6 Joules, both with fresh-faced
propellant bars and after allowing sufficient operating time for the outermost ~1cm of
propellant to ablate. The latter more properly represents the actual operating condition of a
flight thruster. Results of these tests are given in table 5.2.

82
Energy Recession Thrust Error
6 J 0 cm 32.3 µN +/- 1.0 µN
6 J 1 cm 25.0 µN +/- 1.1 µN
5 J 0 cm 25.5 µN +/- 0.6 µN
5 J 1 cm 19.0 µN +/- 0.9 µN
4 J 0 cm 20.2 µN +/- 1.1 µN
4 J 1 cm 12.8 µN +/- 0.6 µN
3 J 0 cm 15.5 µN +/- 1.4 µN
3 J 1 cm 5.6 µN +/- 1.6 µN
Table 5.2: Thrust Stand Performance Study

This test series indicates that the existing thrust stand, when operated in forced harmonic
oscillation mode, is capable of measuring μPPT thrust with a precision of approximately +/- 1 μN,
as compared to the +/- 200 mN precision of the same thrust stand in steady-state mode. It also
indicates that reliable thrust measurements will require first operating the thruster for a burn-in
period of several hours, to ablate propellant to the normal operating geometry from the artificial
flat surface of a fresh bar.

83

Chapter VI

Two-Stage μPPT Performance

The most important aspect of any spacecraft propulsion system’s performance is reliability. Any
shortfall in thrust or specific impulse will result in a reduction of mission capability, but a
propulsion system failure will probably result in a mission failure. This is particularly true for the
μPPT, which is intended for use on small, low-cost spacecraft which will not have great
redundancy. Also critical for many μPPT applications, is precision. If single, small impulse bits
are being used for precision stationkeeping, e.g. in an multi-satellite interferometer system, those
impulse bits must come exactly when expected and be of exactly the desired magnitude.
Thus, once the µPPT design demonstrated operability, the development effort shifted to
demonstrating long–term reliability. The relative simplicity of the design lends itself to the
simplest form of lifetime testing. A complete µPPT system of PPU, capacitors, and propellant
module is assembled, placed within a vacuum chamber, and fired continuously until it fails or
expends all of its available propellant. Testing focused on empirically optimizing the µPPT in
terms of discharge energy, voltage, geometry, and electrode materials. The end result of this
process is a baseline design that has exhibited a high degree of reliability.
6.1 Thruster Life
A total of 12 long–duration tests have been performed on the µPPT baseline design. Of
these 12 tests, two failed within a few hours and the remaining 10 operated successfully for
over 100 hours until the propellant was expended. The two failures were attributed to
propellant fabrication errors. Such failures are not a significant concern since they will be
84
discovered during short–duration functionality testing prior to flight hardware delivery.
Furthermore, since they are attributable to propellant fabrication, the failure rate should
decrease as the fabrication procedures are refined.

In figure 6.1, the propellant has receded within the anode tube about 6” until it resides
slightly to the right of the end of the anode tube, as marked in the photograph. The blue
µPPT plume is still visible at the exit plane of the thruster on the right. The orange visible
emission on the left side of the anode tube is light from the discharge arc that is transmitting
through the short region of the remaining solid propellant. Shortly after the photo in figure
5.16 was acquired, the Teflon™ propellant receded past the left edge of the anode tube, and
the thruster failed due to a full expenditure of propellant. It should be noted that this was an
open-circuit rather than short-circuit failure, did not result in any damage to the power
electronics, and would not have precluded operation of other thrusters using a common
power supply.

Figure 6.1: Two-Stage µPPT Fires Until Propellant Depletion
MicroPPT Plume Propellant Face
Visible Light from Discharge
Transmitting through Teflon
MicroPPT Plume Propellant Face
Visible Light from Discharge
Transmitting through Teflon
85

It has been theorized that using optical means such as a photodetector or photodiode, a rough
estimate of µPPT fuel recession can be done simply by measuring the intensity of the
transmitted light at the rear of the µPPT.

The baseline two-stage µPPT failed only due to propellant exhaustion. Aside from non-
repeatable failures presumed due to manufacturing and assembly defects, only one other
failure mode was observed with the µPPT. At sufficiently low discharge energies, char
formation on the propellant was observed, which initially resulted in erratic discharge
behavior and ultimately in failure due to short circuit through the char between the
electrodes. An example of char formation in a two-stage μPPT is shown in figure 6.2

Figure 6.2: Char Formation in a Two-Stage μPPT

Char formation was observed at discharge energies of 0.64J but not 1.0J in the baseline μPPT
design with 0.141” propellant diameter. In thrusters of 0.250” diameter, char formation was
observed at discharge energies of 3.0J but not at 4.0J. This suggests that the energy threshold for
char formation is approximately 100 milliJoules per square millimeter. This is a somewhat
surprising figure in light of the routine operation of larger PPTs at energy densities of 30-40
86
mJ/mm
2
, and may represent a scaling effect associated with easier dissipation of heat through the
walls of the μPPT. It does establish a critical scaling parameter for μPPT design, in that the
energy density at the propellant face must be maintained safely above 100 mJ/mm

.
6.2 Propellant Consumption
Once long–term functionality had been established, tests were performed to investigate the
potential for changes in the µPPT operation over time. The key issues in this regard are
thrust and propellant consumption rate. Propellant consumption rate over long duration
firings is shown in figure 6.3. For these tests, the three–electrode µPPT was fired at 2.25 J
for at least four hours with a firing frequency of 1.0 ± 0.1 Hz. The thruster was then
removed from the vacuum chamber, allowed to cool, weighed, and placed back in the
vacuum chamber for continued firing. Tests in this manner were conducted on 5 different
samples and extended up to 90 hours. Apart from an initial transient period, the propellant
consumption rate is relatively constant for all of the samples until the propellant is fully
expended.

Figure 6.3: Propellant Consumption Over Long–Duration Firings
0
5
10
15
20
25
0 204060 80 100
Time (Hours)
Propellant Consumption Rate (µg/s)
0
5
10
15
20
25
0 204060 80 100
Time (Hours)
Propellant Consumption Rate (µg/s)
87

The propellant consumption rate in the current design of the µPPT raises two concerns.
First, during the first four hours of testing, one sample is observed to have a propellant
consumption rate three times higher than the average rate. This initial high rate is attributed
to propellant conditioning — the initial coning of the fuel to the concave geometry indicative
of long–term operation.

A second concern from the tests is that the propellant consumption rate is generally higher
than that observed in breech–fed PPTs. The average rate is 8.6 µg/s with a standard
deviation of 1.3 µg/s. By comparison, the propellant consumption rate for the breech–fed
LES 8/9 was 28 µg/s for a 20 J, 1 Hz discharge. Using a linear scaling with energy and
power, the propellant consumption rate for the µPPT is 2.7 times higher than that of LES
8/9.

One contributing cause is in the experimental procedure used to measure the propellant
consumption. In previous PPT designs, the solid block of Teflon™ was simply removed,
weighed, and reinstalled. In the µPPT this is impossible since the Teflon™ propellant is
tightly integrated with the electrodes. The Teflon™ cannot be removed without causing
permanent damage to the propellant module. Therefore, when the propellant consumption is
measured for a µPPT, the measurement necessarily includes the mass loss of the electrodes.
The mass fraction in the propellant module for the intermediate and center electrodes
(ignoring the outer electrode since it largely survives) is 24.5%. If this mass is not counted
in the propellant consumption rate, the average rate falls to 6.6 µg/s, which is still 2.1 times
that expected by a linear scaling to the propellant consumption rate measured on the LES 8/9
88
PPT. It should also be noted that the electrode mass should not necessarily be excluded
when considering the thruster specific impulse. This electrode material participates in the
discharge and is likely accelerated to some degree along with the Teflon™ propellant.
Electrode mass is also ablated and accelerated along with the Teflon™ propellant in
traditional PPTs, but the mass fraction of the metal in the effluent is negligible.

It is also important to note that some other PPTs have had similarly high propellant consumption
rates. For example, the gasdynamic PPT developed by Burton had an average propellant
consumption rate of 40 µg/s for a 10 J, 1 Hz discharge
6,8
This rate is slightly higher than that
observed in the µPPT, and 2.9 times that measured on the LES 8/9 PPT. The Burton thruster was,
like the μPPT, a coaxial device, and was designed to exploit electrothermal as well as
electromagnetic acceleration of the propellant. This result suggests that the μPPT may to some
extent be an electrothermal device as well.
6.3 Discharge Energy Variation
As mentioned earlier, high precision as well as high reliability is required for many μPPT
applications. Because the two-stage μPPT operates in a self-breakdown rather than triggered
discharge mode, the possibility of substantial pulse-to-pulse variation must be considered. As our
thrust stand can not measure single-discharge impulse, this area can most effectively be addressed
by monitoring trigger and main capacitor voltage at the time of discharge. If discharge energy
remains constant, it is probably reasonable to assume that delivered impulse is similarly constant.
Unfortunately, measurement of capacitor voltage an operating μPPT is extremely troublesome, as
the 100 MΩ resistance of the trigger circuit is comparable to the impedance of most commercial
89
high-voltage probes. In order to perform a non-intrusive voltage measurement in this
environment, it was necessary to construct a voltage probe of nearly infinite impedance.
This was accomplished by exploiting the Pockels effect. A Pockels Cell rotates light polarization
in proportion to an applied electric field. To measure voltage, a polarized laser is directed through
the Cell, rotating the beam polarization as shown figure 6.5. A polarized filter attenuates the beam
and the intensity is measured by photodiodes before and after the Pockels Cell and filter
combination. The ratio of the beam intensities is a precise and repeatable function of the applied
voltage. Since the Pockels Cell draws no current, the near–infinite impedance assures a non–
intrusive measurement of capacitor voltage for the µPPT.
Figure 6.4 shows the completed Pockels Cell apparatus. While somewhat cumbersome, it was
adequate for our purposes. With the use of diode rather than He-Ne lasers and a more compact
optical train, it may be possible to reduce the mass and volume of the Pockels Cell system to the
point where it could be used as a diagnostic or performance monitor on a flight µPPT.
‘

Figure 6.4: Pockels Cell Apparatus
90

Figure 6.5: Pockels Cell Setup

The sample dataset included in Figure 6.4 is typical of the results given by this diagnostic. As can
be seen, while there is substantial variation in trigger voltage at discharge, the main discharge is
always at or very near nominal maximum voltage and thus energy. There is little pulse-to-pulse
variation in discharge energy. Pulse interval, however, does vary somewhat. Similar results are
shown in figure 6.6, which shows the discharge voltage only of a μPPT operating at a 2 Hz
discharge rate over a five-second period. At the higher pulse frequency, there is less variation in
pulse interval, which we attribute to their being less opportunity for an anomalously low-voltage
discharge during the steady increase in trigger voltage. Main discharge voltage, and energy, are
again quite uniform.
C
trig
0.01 uF
C
main
= 0.5 uF
500M Ohm
PPU
Laser
BS
Signal
Detector
Polarization
Filter
Ref
Trigger
Voltage
Pockels Cell:
ΔΨ with
applied E Main Voltage
Trigger Voltage
DATA
C
trig
0.01 uF
C
main
= 0.5 uF
500M Ohm
PPU
Laser
BS
Signal
Detector
Polarization
Filter
Ref
Trigger
Voltage
Pockels Cell:
ΔΨ with
applied E
C
trig
0.01 uF
C
main
= 0.5 uF
500M Ohm
PPU
Laser
BS
Signal
Detector
Polarization
Filter
Ref
Trigger
Voltage
Pockels Cell:
ΔΨ with
applied E Main Voltage
Trigger Voltage
DATA
Main Voltage
Trigger Voltage
DATA DATA
91

Figure 6.6: 2.25 J, 2 Hz µPPT Breakdown Voltage

As a final test of long-term repeatability, several μPPT configurations were subject to 100-hour
tests with the Pockle Cell diagnostic in place. Figures 6.7 and 6.8 show the discharge voltage and
energy over one such test, in this case using a 0.250” propellant diameter and 6-Joule discharge
energy. As can be seen, while there are some anomalously low-energy discharges, the vast
majority of discharges fall within 5% of nominal voltage and 10% of nominal energy. Outside of
the initial burn-in period, less than 0.5% of discharges occurred at less than 90% of nominal
energy. Similar results were obtained with other μPPT configurations.

92

Figure 6.7: Breakdown Voltage of a µPPT Fired for 350,000 s at 6 J

Figure 6.8: Pulse Energy of a µPPT Fired for 350,000 s at 6 J
0
1
2
3
4
5
6
7
8
0 50000 100000 150000 200000 250000 300000 350000
Time [s]
Energy [J]
0
1000
2000
3000
4000
5000
6000
7000
0 50000 100000 150000 200000 250000 300000 350000
Time [s]
Voltage [V]
93
6.4 Thrust Measurements
Initial thrust measurements were performed on the baseline two-stage µPPT at a nominal
energy of 2.25 J at ~1 Hz. The average measured thrust was 14.1 µN, corresponding to a
thrust–to–power of 6.3 µN/Watt. While this value exceeded our baseline commitment to
TechSat 21, it was nonetheless discouraging.

In a parallel research effort, the Busek Company of Natick, Massachusetts was given a 0.25”
µPPT to fire on their newly–developed thrust stand. Busek’s stand operates by measuring
the damped sinusoid generated by a single thruster pulse. Given the preliminary nature of
their thrust stand and the small thrust level of the µPPT, the error bars on their measurement
are quite large, but they were able to measure 9 µN/Watt at 4 J, a value consistent with our
results.

Busek’s research suggested that operating at higher capacitance and lower voltage for the
same discharge energy would provide superior performance. They also found that reducing
the inductance of the discharge capacitor also increased performance, not surprising given
that any impedance outside of the thruster head represents an energy loss.

Further testing was therefore confined to two thrusters. First, a modified baseline thruster
using a 2-μF Custom Electronics capacitor at 1500 V, for the standard 2.25 J discharge
energy. Second, a low-impedance thruster using a proprietary capacitor of advanced design,
provided by Busek, still of 2-μF capacitance but with inductance reduced an order of
94
magnitude to ~10 nH. Discharge voltage on this low-inductance µPPT was limited to 1400
V, for a discharge energy of 1.96 J. Both thrusters were operated at a nominal 2 Hz.

Figures 6.9 through 6.12 show measured thrust for these devices over tests conducted to
failure or propellant depletion. While it was not possible to monitor propellant consumption
over the course of the tests, due to the difficulty of cycling Chamber 2, before-and-after
measurements of thruster mass allowed the calculation of average propellant consumption
per discharge. This figure was 16.4 μg/pulse for the 2.25J discharge, 15.3 μg/pulse for the
1.96J discharge, 6.4 μg/pulse for the 1.0J discharge, and 2.9.4 μg/pulse for the 0.64J
discharge. All tests save the second were terminated manually when it appeared that
propellant exhaustion was imminent; the second test (at 1.96J energy) was inadvertently
conducted to actual propellant exhaustion and thruster failure.

Figure 6.9: Modified Baseline µPPT Performance (2.25J, 2 Hz)

200
150
100
50
0
Thrust [µN]
20 15 10 5 0
Elapsed Time [Hours]
1200
1000
800
600
400
200
0
I
sp
[s]
95

200
150
100
50
0
Thrust [µN]
30 25 20 15 10 5 0
Elapsed Time [Hours]
1200
1000
800
600
400
200
0
I
sp
[s]

Figure 6.10: Low-Impedance µPPT performance (1.96J, 2 Hz)

70
60
50
40
30
20
10
0
Thrust [µN]
40 30 20 10 0
Elapsed Time [s]
1000
800
600
400
200
0
I
sp
[s]

Figure 6.11: Low-Impedance µPPT performance (1.00J, 2 Hz)
96

Figure 6.12: Low-Impedance µPPT performance (0.64J, 2 Hz)

Several features are clearly visible in these plots. First, there is a steady decrease in thrust
over the life of the thruster. This is probably due to increasing viscous losses as the
propellant face recedes within the outer electrode. Even the end-of-life performance of the
baseline μPPT is adequate for near-term missions such as TechSat 21, but reducing viscous
losses would certainly enhance the performance of the μPPT by a large degree.
This will be discussed in a later section.

Second, the 1.00J case shows a weak, and the 0.64J case a strong, sinusoidal variation in
thrust during the first half to two-thirds of the test period. This was at first believed to have
been some sort of thrust stand resonance effect. However, the fourth test was paused at the
45-hour mark and the thrust stand recalibrated. The sinusoidal variation in thrust continued
35
30
25
20
15
10
5
0
Thrust [µN]
100 80 60 40 20 0
Elapsed Time [Hours]
1200
1000
800
600
400
200
0
I
sp
[s]
97
where it had left off, with a slight vertical displacement but no change in phase. This was
not a thrust stand effect.

As described earlier, the 0.141” μPPT exhibits surface charring and thus erratic discharge
behavior at energies below 1.0 Joules, and this is now believed to be the cause of the thrust
variation in that test run. Why this behavior manifests as a sinusoidal variation, and why this
variation disappears after ~60 hours of operation, is unknown. Further tests to study this
phenomenon were hindered by a power-supply failure, and rendered largely moot by the
recommendation to operate at discharge energies of 1+ Joules in flight operations..
6.5 Performance Summary
The results of these performance tests are summarized in table 6.1. “BOL” and “EOL” refer
to beginning-of-life and end-of-life performance, respectively. Total impulse, is integrated
over the life of the thruster and assumes full consumption of a propellant bar of 101.6 mm
length. This required modest extrapolation in all but the second case, as the tests were
terminated prior to actual propellant depletion.

Table 6.1: µPPT Overall Performance
Thruster Energy Power Thrust/Isp,
BOL
Thrust/Isp,
EOL
Total
Impulse
Baseline 2.25 J 4.50 Watt 171 µN,
1066 s
47 µN,
293 s
4.80 N–s
Low–I 1.96 J 3.92 Watt 172 µN,
1146 s
38 µN,
253 s
4.85 N–s
Low–I 1.00 J 2.00 Watt 60 µN,
955 s
18 µN,
289 s
5.08 N–s
Low–I 0.64 J 1.28 Watt 36 µN,
1294 s
7.8 µN,
277 s
11.0 N–s

98

While the operation of the low-inductance thruster at 0.64 Joules showed a remarkably high
total impulse, it was also associated with charring and erratic behavior. It is therefore
recommended that flight μPPTs be operated at a discharge energy of 1-2 Joules per pulse,
with pulse rate determined by power availability and/or thrust requirement.

At this energy, a thrust:power ratio of ~9.4 μN/Watt can be expected even at end of life, and
a total propulsive impulse of ~5.0 N-s can be obtained from a single, 0.141” diameter by
4.0” length propellant bar. Operation under these conditions is shown to be highly precise
and reliable.

If a total impulse of greater than 5.0 N-s is required, it may be worth considering the use of
0.250” propellant bars. A μPPT using a 0.250” propellant bar should operate at 4-6 Joules
per pulse and can probably be expected to deliver a thrust:power ratio of 10 μN/Watt and a
total impulse of 25 N-s from a propellant bar of 6” length. Further testing would of course
be required to confirm these figures; such testing was not performed during this project due
to lack of time and sponsor interest.

99

Chapter VII

μPPT Plume Studies

Thruster plumes represent interesting fields of study for three reasons. First, they may offer
direct physical insight into the operation of the thruster. Second, they can provide data for
the validation of numerical models that may in the future support thruster design and
operational activity. Third, the thruster plume represents a source of potentially damaging
interaction with the host spacecraft, through surface erosion or deposition or through bulk
plasma effects.

Numerical modeling of the μPPT, due to the complex physics involved, has proven
extremely difficult. A modeling effort was conducted by the University of Michigan in
conjunction with this effort, resulting in a hybrid fluid–PIC–DSMC model that has had
limited success in predicting electron and neutral densities in the near-field plume region.
This work was supported by a plume diagnostic effort conducted jointly by WER, AFRL,
and the University of Michigan, and while this author’s contribution to the diagnostic effort
was minimal, a summary is presented for completeness and for possible relevance to the
issue of propellant face charring.

Of substantially greater importance, is the issue of plume interactions with the host spacecraft. In
order to at least bound this effect, an observational study of the plume was conducted. The goal
of this effort was to determine whether major plume effects were confined to a well-defined
region, and if so to measure the extent of that region. It is highly unlikely that any spacecraft
100
designer would expose mission-critical surfaces or components to direct PPT plume impingement,
so little effort was made to quantify the magnitude of plume effects where they were present.
7.1 Near-Field Plume Diagnostics
The small plasma volume generated by the µPPT creates a significant diagnostic challenge.
For material probes, such as electrostatic or magnetic field probes, the characteristic length
of the probe is comparable to or larger than the µPPT plasma volume. For interferometric
techniques to measure density, the measurement resolution is constrained by the short scale
length of the plasma. Since the interferometer measures a phase shift proportional to the
product of the density and the laser path length through the plasma, the fundamentally short
path length results in excessive measurement uncertainty for the line–averaged plasma
density.

To address this problem, a Herriott Cell
10
interferometer was used and a technique
employing a ‘point measurement’ was developed
3
. This technique converges multiple laser
passes in a Herriott Cell down to a small area, thereby providing increased laser path length
within the plasma. By focusing a large number of laser beams into a small measurement
volume, the Herriott Cell increases the instrument resolution by a factor of about 10,
compared to a single–pass interferometer, with minimal degradation of spatial resolution.
Electron density is measured with a single–wavelength Herriott Cell and compared with
numerical predictions. Efforts to measure the neutral density after the current pulse using a
second laser frequency are discussed in a previous work.

101
Previous modeling efforts have indicated that the plasma distribution in the plume field
heavily depends upon upstream boundary conditions
4,11
. Therefore the model of the plasma
generation in these devices becomes a very important aspect of accurate plasma plume
simulation. This represents a major difficulty in developing an overall model of μPPT
operation – as plasma generation is driven by radiative ablation of the propellant surface, a
simple upstream boundary condition cannot be provided for the plasma model. Instead, a
coupled plasma and ablation model must be developed.

The composition of the near-field plasma may also be important to the charring observed in
relatively low-energy operation
13,24
. One likely explanation for this process is that an excessive
neutral vapor density in the near-field plume, due to insufficient ionization at low energy, may
result in the deposition of solid carbon on the propellant face. There is as previously mentioned a
substantial neutral vapor component evolved from the propellant surface in the 2-300
microseconds following the discharge, but this occurs at temperatures too low to decompose the
Teflon propellant into elemental carbon and fluorine. Observations or reliable model predictions
of neutral vapor in the near-field plume during the discharge, however, would suggest a carbon
vapor population could be responsible for low-energy charring.
7.2 Herriot Cell Interferometer and Two-Color Interferometer
The µPPT testbed for the Herriott Cell measurements uses a two–electrode, 0.250” µPPT.
The outer diameter of the thrusters tested here is 6.35 mm while the outer Teflon™ diameter
is 5.46 mm. The cathode diameter is 1.64 mm. A DC–DC converter charges a 0.417 µF
capacitor to 5.6 kV, corresponding to a stored energy of 6.6 J. An external LES 8/9
sparkplug and trigger circuit was used to initiate the discharge in the testbed configuration.
102
The plug is placed approximately 2 cm from the propellant face, at a 45–degree angle, and
the discharge is triggered by the data acquisition system.

The interferometer employs quadrature heterodyning technique and a Herriott Cell for
increased path length exposure. Detailed descriptions of these are referenced elsewhere
3
.
Figure 7.1 gives a general layout of the interferometer with the optional second laser
frequency.

Figure 7.1: Interferometer Setup (Two–Color Optional)

A Bragg cell splits the 150 mW Argon–ion laser (488 nm) into scene and reference beams
using a 40 MHz shift in the reference beam. The scene beam is directed into the vacuum
tank through a viewport and into the Herriott Cell optics. The Herriott Cell optics for the
point measurement configuration require the Herriott Cell mirrors as well as focusing and
guiding optics for the input and exit beams. The µPPT is situated halfway between the
Herriott Cell mirrors.

103
The scene beam is returned to the external optics table and recombined with the reference
beam that travels the same path length. The combined beams are focused on the detector for
optimal measurement intensity.

A two–color interferometer was also constructed for baseline comparison of electron and neutral
densities. This interferometer allows separation of electron and neutral densities throughout the
pulse. A 488–nm Ar–ion laser was used simultaneously with an 1152–nm HeNe laser providing
direct measurements of both electron and neutral densities early in the pulse. This data is taken at
5–mm distance from the fuel face and provides the control case for density measurements with the
Herriott Cell.
7.3 Near-Field Density Measurements
The data from the two–color interferometer shows electron and neutral density at 5 mm from
the fuel face for 6.6 J discharge and 13 passes. Figure 7.2 shows this data for the first 30 µs
after the discharge. Peak electron density reaches 2.1 ± 0.3 × 10
16
cm
–3
.

Figure 7.2: Electron Density with Vibrations for 13–pass Herriott Cell
(0.25” Diameter µPPT, 5 mm from Fuel Face, 6.6 J)
104

Over a large number of discharges, uncertainty introduced by vibration effects and shot–to–
shot thruster variation overwhelms the neutral density signal disallowing a definitive
measurement. For a single discharge, the neutral density is typically resolved. Figure 7.3
shows neutral density data with typical error bars taken at 5 mm from the fuel face using the
13–pass Herriott Cell. The data indicate that at 200 µs after the discharge, neutral density is
no larger than 8.5 × 10
16
cm
–3
.

Figure 7.3: Neutral Density Measurements from 13–pass Herriott Cell
(0.25” Diameter µPPT, 5 mm from Fuel Face, 6.6 J)

7.4 Thruster and Model Comparison
A significant finding of this experiment is the degree of agreement between theoretical
modeling predictions and experimental results. A figure of merit is required to compare and
contrast systems that are not similar in energy or dimension. The energy–to–area ratio
105
provides such a basis and is used here to investigate the comparison between past and
current predictions and measurements.

A hybrid fluid–PIC–DSMC approach is used to predict electron density
12
The energy–to–
area ratio for this model is higher than that of the experimental setup used. This is balanced
by increased electron density.

Predictions made for the 0.25” diameter µPPT show strong agreement with experimental
data. Description of the model is referenced elsewhere
4,11
. Figure 7.4 shows direct
comparison of predicted electron density using an experimental current waveform as input.
The red line shows density measured using the Herriott Cell. The degree of agreement is
apparent, and provides support for the model being developed.

Figure 7.4: Predicted vs. Measured Electron Density Time Variation
(0.25” Diameter µPPT, 5 mm from Fuel Face, 6.6 J)
-5
0
5
10
15
20
25
30
35
0 5 10 15 20 25 30
Comparison of 13 pass HC measurement
and Michigan Simulation, 6.6 J, 0.25" MicroPPT
Experiment
Simulation
Time (μs)
106

Herriott Cell measurements give strong evidence of plasma presence as much as 5 mm behind the
exit plane. A µPPT ablation model referenced elsewhere and validated by this data predicts non–
zero carbon ion densities at the exit outside a 3.1 mm diameter µPPT
12
. Incomplete ionization,
neutral backflow and external arc attachment are possible explanations for this phenomenon, and
so may be responsible for propellant face charring.
7.5 Plume Contamination Study
The presence of µPPTs in the vicinity of critical spacecraft systems such as solar arrays,
requires that the plume divergence of the μPPT be quantified to ensure that damage to the
solar array due to μPPT effluent is minimized.

The μPPT is unique in that the effluent is in two stages. The initial stage consists of a high–
velocity, energetic plasma consisting primarily of ionized Carbon and Fluorine, the
constituent elements of the Teflon propellant. This plasma exists for some 10–20 μs after
the thruster fires. This plasma can damage the solar cells via erosion and electromagnetic
interference, although the effect of the latter should be negligible. The secondary stage,
called Late–Time Vaporization (LTV), consists of neutral vapor and microparticles of solid
material that are accelerated out of the thruster as the fuel cools. These particles are
generally 5–50 microns in diameter and exit the thruster from 100–1000 μs after the thruster
fires. Previous research
21
indicates that vapor deposition and particulate impact are spatially
correlated, suggesting that the particles are entrained in the neutral vapor. Both vapor
107
deposition and particle impact can form an opaque film over solar arrays and optical
apertures, degrading performance.

If the particulate and neutral vapor emissions are strongly correlated, plume contamination
of spacecraft systems can be avoided by determining the volume occupied by the plasma
plume and/or by particulate emissions, and designating this a keep-out zone for critical
systems. Alternately, baffle plates can be mounted to block portions of the plume and
prevent adverse interactions.

For this study, a 0.250" diameter μPPT sheathed in copper was fired at 5 J in a vacuum
chamber pumped down to ~20 μTorr and observed with a Cooke DiCam, a gated camera
capable of resolving nanosecond–time scale images. The initial configuration of the μPPT
has the Teflon fuel extending fully to the exit plane of the thruster. As the thruster is fired
on–orbit, the propellant will slowly recess back inside the outer shell. By end–of–life
(EOL), the fuel will be recessed several inches within the µPPT body. Images of both the
plasma and the LTV plume were taken at several fuel depths corresponding to various stages
in μPPT life. Plasma emissions are observed by gating the camera to observe the first 20 μs
of the discharge, while LTV is observed over a period from 20 μs to 200 μs post-discharge

Noticeable in most of the following pictures is a curved rod about 1 cm in front of the thruster.
This is an external spark igniter used only in these tests, as operation in a self-breakdown mode
would preclude synchronization of the high-speed camera with the thruster. It should also be
noted that in all cases gain was adjusted to produce clearly visible images regardless of source
108
intensity. The relative brightness of different images should not be taken as an indication of actual
plume intensity.
7.6 Plume Contamination Results
Images of the plume resolved by the DiCam were uniform and repeatable at all propellant depths.
LTV images, by their very nature, are erratic from shot–to–shot as they capture trails of individual
particles exiting from the thruster rather than a uniform, expanding plasma. The neutral vapor
emissions cannot be directly observed, and are assumed to be spatially correlated with the
particulate emissions. Despite this irreproducibility, the overall divergence of the LTV was
repeatable at all depths.
Figure 7.5 shows the μPPT plume at BOL, when the Teflon propellant is flush against the
exit plane of the copper tube. The plasma plume is initially well–directed with virtually no
divergence, a characteristic common at all depths. Further into the flow field, the plasma
diffuses into a 60° cone.

Figure 7.5: BOL Plume, No Recession

109
LTV at BOL, displayed in Figure 7.6, progresses into the complete 180° hemisphere. Images
at increased fuel depth indicate a progressive decrease in the LTV divergence angle. This
implies that the presence of a nonconductive baffle or nozzle at the exit plane would serve to
collimate the LTV to reduce Teflon deposition on critical systems.

Figure 7.6: BOL LTV, No Recession

The fuel was recessed 7 mm for images presented in figures 7.7 and 7.8. The plasma plume
diverges into a 55° cone – no appreciable change from the BOL case.

Figure 7.7: 7 mm Recession Plasma

110

Figure 7.8: 7 mm Recession LTV
The LTV is somewhat improved, compressing to 120°. A rough geometrical model for LTV
divergence is displayed in Figure 7.9. Using this model with the 7 mm recession depth and
6.35mm thruster diameter yields a divergence of 84°. That the measured divergence angle
exceeds that predicted by the geometric model, seems to confirm that the particulate
component of the emission is entrained in an expanding neutral vapor phase.

α

Figure 7.9: Geometrical LTV Divergence Model

Figures 7.10 through 7.13 show results at recession depths of 17.5mm to 54mm. The plasma
plume maintains a relatively constant 60
o
divergence, while the LTV emission becomes
progressively narrower. In all cases, observed LTV divergence is 30-50% broader than
predicted by the geometric model.
111

Figure 7.10: 17.5 mm Recession Plasma

Figure 7.11: 17.5 mm Recession LTV

Figure 7.12: 54 mm Recession Plasma

112

Figure 7.13: 54 mm Recession LTV

A summary of the measured divergence angle, as a function of the recession depth of the
propellant is shown in figure 7.14 for both the plasma and LTV (neutrals/particles) emission.
The images above and the plot of figure 7.14 suggest that intentionally recessing the
propellant face by 1-2 cm at beginning of life may significantly reduce the potential for
spacecraft contamination from the neutrals and particles. If this is done, dangerous plume
conditions will be confined to a roughly 60
o
cone centered on the thrust axis.

Figure 7.14: Divergences of the Plasma and LTV Plumes
0
45
90
135
180
0 204060
Propellant Depth [mm]
Divergence Angle
Neutrals/Particles
Plasma
0
45
90
135
180
0 204060
Propellant Depth [mm]
Divergence Angle
Neutrals/Particles
Plasma
113

Chapter VIII

Advanced μPPT Concepts

Several limitations were discovered during the previously-described research which, while not
precluding the use of the μPPT in the TechSat 21 and other near-term missions, must be addressed
if the thruster is to have broader applicability. Most critical of these are, the relatively low thrust
to power ratio, the limited total impulse of the largest practical propellant bar, and the requirement
for a separate power supply and capacitor for each thruster. Possible approaches have been
identified for overcoming each of these limitations, and preliminary experiments were performed.
None of these resulted in immediate success, but all did suggest areas for future research.
8.1 Chemically Energetic Propellant
As mentioned in section 4.5, a chemically energetic propellant suitable for PPT operation
was identified and tested. This proved to be at least functional, which few other alternate
propellants can claim, but offered no performance gains in normal PPT operation. This is
likely due to the fact that chemical augmentation can only produce thrust by gasdynamic
operation, which conventional PPT designs cannot readily exploit. However, results
presented in section 6.2 suggest that, especially late in life when the propellant face has
receded deeply into the electrode structure, the μPPT does produce a substantial amount of
its thrust via gasdynamic acceleration. The use of a chemically energetic propellant should
be reevaluated for the μPPT.

The principal difficulty in accomplishing this, is that commercial coaxial cable cannot be
used to assemble the testbed μPPT. With some difficulty, a jig was developed which
114
allowed the assembly of a μPPT propellant/electrode assembly from a length of narrow
copper tubing, copper wire, and a supply of melt-castable, non-conductive chemically
energetic propellant. However, this procedure could not produce the three-electrode
geometry necessary for a two-stage μPPT.

Preliminary tests were conducted using a triggered, single-stage μPPT and chemically energetic
propellant, before the sponsor of this research effort directed that work be focused on more
immediately practical applications. These tests demonstrated basic functionality and reliable
operation, but did not extend to performance measurements. This remains a potentially fruitful
area for future work.
8.2 Thinned Outer Electrodes
The outer electrode of the baseline μPPT is copper of approximately 0.022” thickness. This
may be excessive in two respects. First, it represents a majority of the mass of the
propellant/electrode assembly, but does not directly contribute to thrust generation. Any
electrode mass in excess of that required to keep ohmic heating in the electrode structure
small compared to other efficiency losses, is wasted. Second, the thickness of the outer
electrode is sufficient to entirely prevent it from ablating. As a result, the propellant face
recesses inside the electrode structure, resulting in substantial viscous losses. While
excessive ablation of the outer electrode would eventually result in thruster failure, ablation
at approximately the recession rate of the propellant could well prevent the thrust decrease
observed over the life of the system.

115
To test this theory, several two-electrode μPPTs were subject to chemical etching to reduce
the outer wall thickness. This was a three-step process:

1. For removing the smallest nicks and general smoothing of the metal surface,
immersion in a mild solution (0.5 molar) of Ammonium Hydroxide (NH
3
OH) was
used. This required several days. The fact that it quickly became saturated with
copper assured us that no harm would come to the tube by leaving in the solution too
long.

2. For removing larger nicks and producing modest thinning of the copper electrode, a
strong solution (3.0 molar) of Nitric Acid (HNO
3
) was used. While working much
faster, this method resulted in noxious fumes and had to be used in a fume hood.

3. If extensive thinning was desired, electrolysis was then used. By attaching the
negative end of a 1.2 volt power supply to the piece and immersing that piece in a
concentrated solution of Copper Nitrate (Cu(NO
3
)
2
). All the copper would migrate
to the positive end of the power supply which was attached to a copper plate, also in
the solution.

Chemical etching requires a uniform solution of a constant temperature and concentration
along with constant agitation. The initial wall thickness of the copper was .022 inches.
Using our removal techniques, wall thicknesses of 0.020”, 0.014”, 0.010”, and 0.007” were
achieved. As the wall thickness reduced it was increasingly difficult to maintain a uniform
thickness, but it was possible. A photograph of a µPPT with outer electrode thinned to
116
0.010” is given in figure 8.1. The resulting µPPT proved functional at all of these wall
thicknesses, but no significant outer electrode ablation was observed.

Figure 8.1: Standard (Left) vs. Thinned–wall(Right) µPPT

As a testbed unit, we were able to produce a µPPT with a wall thickness of 0.003” along its
first inch. It was hoped that with extremely thin walls the outer electrode would ablate way
during use. Instead, small thin pieces of the copper would stay attached to the outer
electrode. While the thruster was firing, thermal expansion and plasma pressure were
sufficient to keep these pieces away from the intermediate electrode. When the thruster
stopped firing the pieces would collapse inwards, cool, harden and make contact with the
intermediate electrode, resulting in a short–circuited µPPT.

It does not appear to be practical to cause controlled ablation of the copper outer electrode.
However, substantial mass savings can be achieved by moderate, controlled thinning of the
outer electrode. Alternately, it might be easier to simply manufacture the electrode from
thinner materials in the first place, though that would preclude the use of readily available
commercial materials and coaxial-cable manufacturing capability.
117

One further area of possible research would be the use of alternate outer electrode materials.
Some early µPPT was conducted using aluminum outer electrodes, and it may be that a
moderately-thinned aluminum electrode would exhibit controlled mass loss (and, of course,
substantial mass savings). However, chemical or electrochemical techniques for etching
aluminum were not readily available in the context of the present research.
8.3 Clustered μPPT
An obvious method considered for increasing the amount of propellant to be carried by the
µPPT is to use multiple tubes in parallel. The self–triggered, three–electrode µPPT is an
ideal candidate for such a modification. Putting two tubes in parallel does nothing to stop the
self–triggering process. Figure 8.2 shows an example of a functional clustered µPPT.

Figure 8.2: Cluster of Two-Stage µPPT Tubes
118

No provision is made for actively controlling which of the seven propellant/electrode
assemblies will self-trigger for any particular discharge. This raises the possibility that some
slight manufacturing defect or other difference will result in one tube discharging
preferentially, depleting its propellant before any other thruster has a chance to discharge.
This would not be a significant problem, as propellant depletion causes a two-stage μPPT to
fail in an open-circuit mode which would force subsequent discharges to occur in other,
undepleted tubes.

However, functional testing of a seven-tube cluster indicates that the discharges are
distributed in a random or nearly random fashion. This results in uniform or nearly uniform
recession, as shown in figure 8.3. Similar results are expected for larger clusters, which
would allow an arbitrary increase in propellant throughput and so total input while allowing
the use of a single power supply and capacitor.

Figure 8.3: Recessed Propellant in a Cluster of µPPT tubes

119

Figure 8.2 shows charring across the propellant face in all seven tubes. When this
configuration was fired it was discovered that after the primary discharge in one tube, there
was sufficient charge in the capacitor to form a second arc across a different tube or tubes.
Since the total energy available was discharged across a much larger surface area (all the
tubes that formed an arc across their main propellant) the energy density was not sufficient
to prevent carbon deposition on the propellant face. System failure due to charring occurred
after less than an hour of operation

Increasing the available energy to reduce charring would work when there were only two or
three tubes. A seven-tube configuration required a five or six fold increase in energy and a
corresponding increase in power supply and capacitor sizes. This would rapidly become
impractical on grounds of excessive mass.

A more practical solution to sympathetic firing of adjacent tubes was to move the trigger arc
further away from the other propellant faces. The results shown in figure 8.3 were achieved
by fabricating the μPPT with the propellant already recessed by approximately 8 mm. This
provided sufficient isolation between each discharge and the propellant faces of adjacent
tubes, that sympathetic discharges were prevented.

This µPPT successfully fired for eight hours before the test was manually terminated. The
extra fuel of six additional tubes more than made up for the lost of ~8mm of propellant from
each tube, and as discussed in section 7.5 such recession is already considered desirable for
the purpose of suppressing plume contamination.
120
Furthermore, the addition of multiple tubes in parallel does not change the electrical characteristics
of the μPPT during discharge, for the simple reason that there is negligible current flow through
the non-discharging tubes. Indeed, with an arbitrarily large number of tubes the innate
capacitance of the other tubes might provide additional energy storage, though this effect is
expected to be minor for systems of practical size. There appears to be no obstacle to the
clustering of up to several dozen μPPT propellant/electrode assemblies in order to increase
propellant throughput and so total impulse.
8.4 Multi-Axis Switching of the μPPT
Simple clustering of μPPT propellant tubes, as described in the previous section, requires
that all tubes be aligned along a common thrust axis in order to provide a predictable thrust
vector. The most likely application of the μPPT is for use as an attitude control actuator,
which would require a minimum of six and more likely twelve to sixteen discrete thrusters.
With the present μPPT design, this would require a similar number of separate power
supplies and discharge capacitors, a significant mass and complexity penalty.

It would thus clearly be desirable to connect multiple μPPT propellant/electrode assemblies,
or parallel clusters thereof, to a single power supply and discharge capacitor with the ability
to selectively fire any particular desired thruster. With the two-stage thruster, this can be
accomplished without having to switch even the modestly high currents of the trigger
discharge – each thrust axis (whether one or several tubes) can be provided with its own
small trigger capacitor, and only the low-current charging of the trigger capacitor need be
switched.
121

A general schematic of the trigger circuit for three thruster elements using a common
capacitor and PPU is given in figure 8.4. As can be seen, the circuit is still quite simple and
only a single TTL input is required to activate the desired thruster element.

Figure 8.4: Multiple Thrust Axis Switching

In this circuit diagram, switching is accomplished by means of a high–voltage SCR. This is
an oversimplification, as the minimum acceptable trigger voltage is well in excess of the
rated holdoff voltage of any space–qualifiable SCR. Furthermore, high–voltage ringing
from the breakdown would propagate through the gate of the SCR and destroy any TTL
logic controlling the switching.

In order to provide robust, survivable high–voltage switching, it is necessary to replace the
single SCR in the conceptual design with a system including multiple SCRs in series to hold
off the necessary voltage, with optical isolation of the trigger input. After some research, the
International Rectifier IRHY7G30CMSE rectifier was identified for this application. This is
+TTL +TTL
122
compact, 1 kV, 4.8 Amp SCR hardened to 100 kilorads and suitable for flight applications.
Presuming a 2 kV trigger voltage remains acceptable; two such rectifiers in series would be
required for each µPPT thrust element. The MII 66099 radiation–hardened optocoupler was
determined to provide suitable isolation and trigger voltage for the gate input of the SCR.

Circuit design was somewhat complicated by the requirement of ensuring simultaneous
triggering of both SCRs. Were one switch to close even fractionally before the other, the
SCR remaining open would find itself holding off the full trigger discharge voltage and thus
immediately fail. After a great deal of experimentation, a circuit was developed in which
only the bottom SCR of the series is directly triggered by the optocoupler, with the voltage
drop across the now–closed switch pulling the base voltage of the upper SCR below the
initial gate voltage and immediately triggering that SCR. The gate voltage of the upper SCR
is delivered by a high–impedance voltage divider coupled to the output, ensuring that the
trigger discharge voltage is uniformly distributed across the SCRs throughout the operation.

A schematic of this circuit is given in figure 8.5.

123

470 Ω
470 Ω
ΜΩ
Ω
Ω
ΜΩ
Ω

Figure 8.5: Optically Isolated High–Voltage Switch

This circuit was modeled using PSpice and later tested experimentally using low–cost COTS
equivalents of the rad–hard components identified above. Simulated and experimental
(single shot) results are given in figure 8.6 for a case with 1 kV applied voltage and an
output load with the impedance characteristics of a µPPT coaxial trigger assembly.

124

0
500
1000
1500
2000
01 23 45 67 89 10
Time, ∪ s
Voltage, V
Simulated
Experimental

Figure 8.6: Switched µPPT Performance

As can be seen, there is a great deal of ringing in the early period of the switching event, due
to the impedance of the simulated µPPT at the output. This prevented us from extensively
testing the circuit at the 2 kV design point, due to overvoltage failures of the SCR in the
early ringing period. However, the simulated and experimental data are in fairly good
agreement as to both the magnitude of the initial bounce and the steady–state voltage
achieved.

The high voltage achieved during the early ringing period suggested that it might be
desirable to trigger the µPPT at this point, rather than with steady–state trigger voltage.
Experiments to this end, using the demonstration circuit operated at 1 kV applied and with
an actual µPPT load, demonstrated approximately 95% triggering reliability, which was
125
promising but insufficient for mission requirements. Efforts to improve the reliability of this
triggering mode were unsuccessful.

Triggering from the steady–state applied voltage, as originally intended, requires
suppressing the internal ringing of the switch circuit so that the steady–state voltage closely
approximates th 2 kV maximum rated voltage of the dual–SCR circuit. This can be
accomplished by applying an RC filter between the switch circuit and the µPPT trigger
electrode. We were unable to complete this work before the termination of the μPPT
development effort at AFRL, but a similar switching circuit was developed and incorporated
in the three-axis Busek flight μPPT described in the next section

126

Chapter IX

The μPPT Flight Unit

Chapter 3 described the selection of the μPPT as the yaw control actuator for the TechSat 21
space flight, which demonstrated the utility of the μPPT and motivated most of the research
just presented. In parallel with this research, a flight μPPT was constructed, tested, and
delivered to carry out the mission and demonstrate the technology in an operational
environment.

However, during the course of this effort, the TechSat 21 spacecraft was redesigned several times,
resulting in corresponding changes to the attitude control system and the PPT. This chapter
describes this process and the resulting flight μPPT.
9.1 TechSat 21 First Redesign
The original concept for the TechSat 21 spacecraft, as described in Chapter 3 ,included a
deployable gravity gradient boom as shown in figure 3.1. With this geometry, propulsive
attitude control would be required only for yaw steering, and the µPPT design developed in
this effort was optimized for this application

In November 2000, the spacecraft bus contractor imposed a major redesign of the vehicle, to
the more conventional three–axis stabilized configuration shown in figure 9.1. This greatly
increased the demand on the µPPT system. Three–axis attitude control increases the number
of thrusters required from six to twelve and required positioning at widely separated
127
locations on the spacecraft bus. Furthermore, the requirement for pitch and roll steering
involves the much higher I
xx
and I
yy
moments of inertia, where yaw steering only involved
rotation about the I
zz
axis.

Figure 9.1: TechSat 21 First Redesign

An improved–capability µPPT system was designed to meet this requirement. To provide
three–axis steering, four µPPT modules were to be used, positioned at the outer corners of
the deployed solar array, with each module containing three propellant tubes oriented along
three thrust axes. Two power processing unit modules were positioned at the outer edge
center of the solar arrays, each feeding two µPPT modules. A single central PPU was
considered and rejected due to the impracticality of deploying high–voltage cables along the
full length of the solar array. This design is shown in figure 9.2

128
ξPPT module
PPU Module
125 cm
ξPPT module
3 cm
8 cm
Power Input (28V Bus)
Telemetry Outputs (4 x 0-5V)
Command Inputs (4 x TTL)
10 cm
High-Voltage Coaxial Output (x2)
Command Inputs (8 x TTL)
High-Voltage Coaxial Input
7.5 cm
Thruster (x3)
Telemetry Outputs (8 x 0-5V
10 cm
Capacitor & Switch Module
2.0 cm 2.5 cm
5.0 cm

Figure 9.2: µPPT First Redesign

This µPPT design retained the operational simplicity of the original concept. Each PPU
module requires only unregulated 28V bus power and, with the requirement to switch
between multiple thrust axes, six TTL command lines associated with individual thrusters.
Six analog telemetry points are returned to the spacecraft, corresponding to temperature and
operating voltage measurements in each of the three modules. All spacecraft interface is
through the PPU module, and thermal, electromagnetic, and other environmental impacts are
minimized. A functional diagram of this system is given in figure 9.3

129

Power Supply
Module
Thruster
Module
Thruster
Module
Propellant
Propellant
Propellant Propellant
Propellant
Propellant
4
4
6
2 2
6
28V Bus Power
2kV Pulse Discharge
2kV Low Current
TTL Command
5V Analog Telemetry
75⎠⎧N Thrust
Micro PPT System
2 per Spacecraft
System Inputs:
28V, 550mA Bus Power (On)
28V, 50mA Bus Power (Off)
6 TTL Command Inputs
System Outputs:
75 mN Total Thrust
10 ng/s Plasma Exhaust
6 0-5V Analog Telemetry Lines
Key:

Figure 9.3: µPPT Functional Diagram

While the spacecraft interface remains simple, the added functionality substantially
increased the mass and internal complexity of the µPPT system. A mass breakdown for the
redesigned µPPT system is given in table 9.1. As can be seen, the target mass of 1 kg was
substantially exceeded. Several mass reduction strategies were explored, mostly involving
component substitutions. These could have resulted in mass savings of 500-600 grams, but
would have required substantial additional testing.

130

2 kV DC-DC High Voltage Power Supply 80 gm x1 80 gm
PPU Electronics & Internal Wiring 40 gm x1 40 gm
PPU Structure & Mounting Bracket 60 gm x1 60 gm
PPU Thermal Control Surface 20 gm x1 20 gm
PPU Total Mass (2 per S/C) 200 gm
0.5 mF, 2kV Capacitor 75 gm x2 150 gm
Switching Electronics & Wiring 20 gm x1 20 gm
Thruster Module Structure & Mounts 40 gm x1 40 gm
Thruster Module Thermal Control 10 gm x1 10 gm
0.25" x 3" Propellant Modules 2.5 gm x12 30 gm
Thruster Module Total Mass (4 per S/C) 250 gm
2kV Coaxial Cable, 0.3m Length 10 gm x2 20 gm
Coaxial Connectors (SMA) 10 gm x4 40 gm
6-Wire Ribbon Cable, 0.3m Length 2.5 gm x2 5 gm
6-Wire Connector 5 gm x4 20 gm
12-Wire Ribbon Cable, 4m Length 60 gm x1 60 gm
12-Wire Connector 15 gm x2 30 gm
28V, 1A Power Cable, 4m Length 40 gm x1 40 gm
28V Connector 5 gm x2 10 gm
Wiring Harness Total Mass (2 per S/C) 225 gm
Micro PPT System Total Mass 1850 gm

Table 9.1: Mass Breakdown for TechSat 21 First Redesign µPPT System

9.2 New Performance Requirements

The redesign of TechSat 21 resulted in a substantial increase in the requirement for ACS
thrust and total impulse, which could exceed the original specifications of the µPPT system.
The detailed ACS requirements of the new TechSat 21 configuration were analyzed to
determine the impact on the µPPT system design.

The TechSat 21 ACS requirements were dominated by 292 cross–track maneuvers for
stationkeeping and formation change. Each of these maneuvers would require two slew
maneuvers in order to align the hall thruster system, for a total of 584 slew maneuvers.

The propellant required for a slew maneuver depended on the spacecraft mass properties, the
magnitude of the slew maneuver, the time allowed for the slew maneuver, and the
131
contribution of the magnetic torquer system to the maneuver. As this was to be a worst–case
estimate, we assumed the magnetic torquers make no contribution, and that all slew
maneuvers would be 90–degree rotations of the spacecraft. A slew time of 30 minutes was
assumed, and the most pessimistic estimate of the I
zz
moment of inertia, 180 kg–m^2, was
used.

The slew maneuver was considered as a constant angular acceleration through the first half
of the required slew, followed by a similar constant deceleration which brought the vehicle
to rest at the completion of the slew. Thus, the µPPT system was required to provide enough
torque for a constant–acceleration slew to 45 degrees (0.79 radians) in 15 minutes (900
seconds). This required an angular acceleration A of:

A = 2 * 0.79 / (900^2) = 1.95E–6 radians per second.

With a 180 kg–m^2 moment of inertia, this required a torque T of

T – 1.95E–6 * 180 = 3.51E–4 Newton–meters.

The moment arm available for the thrusters was 4 meters, and the thrust axis was oriented 45
degrees away from optimum for Z–axis slews, so the total µPPT thrust required was

F = 3.51E–4 / 4.0 cos 45 = 1.25E–4 Newtons

132
This could be provided by two opposing µPPT thruster elements at a thrust of 62.5
micronewtons, close to the 75 micronewton nominal maximum thrust value. After fifteen
minutes, the spacecraft would have rotated through 45 degrees and would have an angular
velocity of 0.1 degrees/second. At this point, a second pair of thrusters would be fired to
counter this rotation, bringing the spacecraft to a halt after another fifteen minutes at a total
angular displacement of 90 degrees.

Thus, a single slew maneuver would involve four thrusters firing at a thrust of 62.5
micronewtons for a period of 15 minutes each, or a total impulse I of

I = 4 * 62.5E–6 * 900 = 0.225 Newton–seconds.

292 cross–track maneuvers with two slew maneuvers each would require a total impulse of:

I = 292 * 2 * 0.225 = 131.4 Newton–seconds

No individual thruster would have to provide more than 32.9 N–s of impulse

As determined in Chapter 6, a baseline µPPT with a single 0.141” propellant bar can deliver only
5 N-s of impulse. However, the use of a 0.250” propellant bar allows a total impulse of 25 N-s,
and clustering two such bars as described in section 8.3 would provide acceptable performance for
this mission.

133
9.3 TechSat 21 Second Redesign
In early 2001, further redesign of the TechSat 21 spacecraft dramatically altered the role of
the µPPT system. While the concept and geometry were largely unchanged, the mass
increased from 137.7 to 170 kilograms, with proportional increases in the principal moments
of inertia. Furthermore, spacecraft operational requirements were found to require a five–
minute slew, rather than the fifteen minute period previously assumed. These changes
collectively increased the torque requirement on the ACS actuators by nearly a factor of
four, substantially exceeding the capability of the µPPT system.

At this point, the bus designer made the decision to demanifest the µPPT as an ACS
actuator. Instead, a set of reaction wheels and magnetic torquers were to be used for this
purpose. Reaction wheels had previously been excluded from consideration due to excessive
mass, but mass constraints were clearly no longer a major concern and the new bus design
could easily support a minimal reaction wheel system.

However, the demonstration of technologies critical to future low–mass microsatellites was
still a major mission of the TechSat 21 program, and the decision was made to fly a single
µPPT thruster as a test unit. This thruster would be mounted on a body panel with the main
propulsion system and several diagnostic instruments, and so would have almost no moment
arm through which to exercise ACS authority. This new configuration is illustrated in
Figure 9.4.

134

Figure 9.4: TechSat 21 Second Redesign

With this redesign, the mass allowed for the µPPT system was reduced dramatically, to not
more than 640 grams. An attempt was made to demonstrate all of the functionality of the
proposed operational system by flying a switchable two–axis thruster module, but the
lightest credible designs for this configuration still exceeded 800 grams. Therefore, our
research effort was directed away from switchable multi–axis systems to a single–axis
demonstration unit using established components and configurations and optimized for
minimum impact on the host spacecraft. A suitable demonstration module was designed and
tested, and is illustrated in figure 9.5.

135

Figure 9.5: µPPT Demonstration Unit

This unit reverts to the original 0.141” propellant module, providing a minimum thrust of 20
micronewtons and a total impulse of at approximately 5 Newton–seconds. Total mass
would be between 390 grams and 550 grams, depending primarily on the final selection of
the potting compound to be used to prevent internal electrical breakdown. Preliminary tests
had indicated that the low–density 20–3035 urethane epoxy had adequate electrical and
mechanical properties for this application, and that a total mass of <400 grams could thus be
achieved, but concerns over its low thermal conductivity would require a full thermal–
vacuum test to address.

The primary design goals of the demonstrator unit were to prove µPPT functionality in the
space environment and to minimize impact on the host spacecraft. The use of 28V bus
136
power, TTL command, and minimal analog telemetry minimized power and C&DH impacts
on the bus; primary environmental impacts were expected to be thermal load, plume
contamination, and EMI. The total heat dissipation of the unit was less than 20 Watts, and
passive thermal control would easily keep the unit within acceptable temperature limits
without posing an unacceptable burden on the bus. However, a localized 8.6 W heat load at
the voltage regulator could have posed an internal thermal failure point for the demonstrator
module itself. Detailed thermal modeling and/or thermal–vacuum testing would have been
required to address this issue, and as mentioned above the thermal conductivity of the
potting compound was a critical undetermined parameter.

Both plume impingement and EMI concerns were addressed through shielding. The µPPT
system was enclosed in a sealed metal–walled box pierced only by the propellant module,
two DC bus power lines, and eight optically isolated command and telemetry lines. The box
was grounded to the spacecraft chassis, while separate power and command/telemetry
ground points were provided to accommodate any desired spacecraft grounding philosophy.
Furthermore, an electrically isolated aluminum shield was extended around and well beyond
the exit plane of the thruster, to physically confine plume debris and electrically shield RF
emissions from the discharge itself. A block diagram of this system is given in figure 9.6.

137

PPU
C&DH
Board
Main Capacitor
Trigger Capacitor
Propellant Module
+28 V
Command
(4 x TTL)
Telemetry
(4 x 0-5V)
Chassis
Voltage Telemetry Leads
Housing Isolated
from Chassis

Figure 9.6: µPPT Demonstrator Unit Block Diagram

This system is essentially a single–axis version of the proposed three–axis ACS actuator µPPT,
with the PPU collocated with the thruster and with the switching electronics removed. Additional
command and telemetry capability is provided to support the technology demonstration mission.
The functionality of all components and of comparable integrated systems has been demonstrated
during the development of the full–capability µPPT system for the earlier TechSat 21 design.
9.4 FalconSat 3
The µPPT demonstrator unit developed for the final version of the TechSat 21 spacecraft,
due to its low mass and minimal spacecraft impact, can be easily adapted to any other small
satellite. In 2001, we were approached by the U.S. Air Force Academy (USAFA) regarding
the possibility of using four single–axis µPPT demonstrator units to provide a minimal
attitude control capability for their student–built FalconSat 3 spacecraft. This is a very small
satellite that would otherwise be completely unstabilized, and it was determined that the
FalconSat bus could support the µPPT system and that the µPPT system could provide basic,
if inefficient, three–axis stabilization for the bus.
138

As program delays associated with the two redesigns had postponed the TechSat 21 mission
well beyond 2005, it was decided that the µPPT demonstrator would have its first flight
opportunity on the FalconSat 3 mission. USAFA’s student spacecraft projects necessarily
adhere to a rigid four–year schedule, and are more tolerant of technical risk than the TechSat
21 program. They thus represent a more suitable environment for early demonstration of
new technologies.

The finished flight–like product, pictured in figure 9.7, was taken to the USAFA vibration–
and shock–testing facility to verify that it could endure the rigors of launch and fairing
separation. The µPPT was subjected to vibe levels equivalent to the Delta IV ESPA level as
well as a 100G separation shock. A post–test inspection indicated that the critical
components survived and functionality was maintained.

Figure 9.7: Flight–Like µPPT on USAFA Vibe Table

139
9.5 Delivery of the Flight Unit
In 2003, several USAF satellite programs were reorganized. One consequence of this was
the demise of TechSat 21 as an independent program. Its mission was folded in with that of
several other programs, to become the TacSat 2 spacecraft. This resulted in still greater
mass increase of the spacecraft, a much tighter mass budget in spite of the allowed mass
increase, and thus the complete demanifesting of the demonstration μPPT from the
spacecraft.

There was, however, a beneficial result of this cancellation. The USAF’s support for a μPPT
flight demonstration was now focused entirely on the FalconSat 3 spacecraft, with additional
resources being devoted to that effort and no further requirement that the μPPT development
team at AFRL divide its efforts between the two projects.

Throughout this research effort, the Busek Corporation had expressed interest in the
commercial manufacture of a μPPT system. In addition to its own in-house efforts, Busek
had collaborated with AFRL and WER to a limited extent in the research described here.
With the Air Force’s decision to focus all μPPT development on the FalconSat 3 project, and
in light of Busek’s greater experience in flight hardware manufacture, a decision was made
to license the WER μPPT to Busek, and have the Air Force Academy contract with Busek to
produce the flight unit for FalconSat 3.

140
Key details of the Busek-produced thruster are of necessity proprietary in nature and will
thus not be discussed here. However, the Busek system consists of a single PPU and
capacitor driving three separate thrust axes, similar to the system described in section 8.4
and proposed for use on the first redesign of TechSat 21 in section 9.1. A photograph of this
system is shown in figure 9.8. The Busek flight prototype μPPT was formally delivered to
the United States Air Force Academy on February 28 of this year, has been integrated with
the spacecraft, and is planned for launch in early 2007.

Figure 9.8: Busek Prototype Flight μPPT

141

Chapter X

Conclusions

While the conventional PPT may represent a technological dead end, the micro pulsed plasma
thruster or μPPT is a promising new propulsive technology for small spacecraft application. A
detailed survey of available micropropulsion systems suggested that the low-powered Hall Effect
Thruster outperformed the conventional PPT in almost all respects for near-term microsatellite
missions such as TechSat 21. The μPPT, however, represented the smallest and simplest
propulsion system likely to be available in the future, and so in spite of its limited performance is
an enabling technology for small microsatellites. Fundamentally, the greatest increase in a
spacecraft’s propulsive capability comes when any propulsion system, however limited, is
incorporated in a satellite that would otherwise be left to drift and tumble.
Three fundamental causes of inefficiency in the conventional PPT design were identified. First,
the inefficient magnetic field geometry requires a great deal of energy be used to generate
magnetic field volume that is not exploited propulsively. Second, the pulsed operation of the
thruster is associated with irreversible transient losses. Third, late-term vaporization and
particulate emission result in 80+% of a PPT’s propellant being ejected at very low velocity, with
negligible propulsive effect. As a result of these issues, PPTs are less than 10% efficient at
converting electric power input into thrust power output.
The inefficient magnetic field geometry could be overcome by using a coaxial thruster design.
Several changes to the thruster pulse discharge circuit and mode of operation were tested in an
attempt to deal with transient loss terms, but these depended on semiconductor switches and
diodes that could not withstand the extreme voltage transients during the discharge. Several
142
alternate propellants were identified that might be less susceptible to late-term vaporization and
particulate emission, but none proved suitable for PPT operations. In most cases, these alternate
propellants failed by surface charring. The XPPT-48, exploiting all of these techniques as best
could be arranged, offered only slight gains in performance over more conventional designs.
The same limitations apply to the μPPT, but as the μPPT is not called upon to deliver great
propulsive efficiency, are not crippling. The major limitation on the μPPT is the small total
impulse available, due to the propellant supply consisting of a single, short, narrow bar.
Increasing the length of the bar would result in excessive viscous losses internal to the thruster.
Increasing the diameter of the bar would lead to excessive system voltage requirements.
Increasing the number of propellant bars would require a complex switching system, eliminating
the major advantage of the μPPT.
The development of the two-stage μPPT overcomes this limitation. The two-stage μPPT consists
of nestled coaxial electrodes and propellant tubes, with a main discharge stage providing most of
the propulsive impulse and a small trigger stage passively coupled to the main discharge. This
allows the main discharge electrode, and propellant bar, to be of arbitrarily large diameter. It also
allows for parallel clustering of multiple tubes with only passive coupling, and active switching of
multiple thrust elements, where desired, can be conducted on the low-current charging side of the
trigger stage only.
Testing of the two-stage μPPT represented a significant challenge, and required the development
of a new type of thrust stand. This work, however, confirmed that the μPPT was a robust, reliable
system with sufficient performance for anticipated missions. It also verified that, by the simple
143
expedient of recessing the propellant face prior to flight, adverse plume interactions with the host
spacecraft can be prevented.
Preliminary experiments were conducted to demonstrate the suitability of high-energy propellants
for the μPPT, and to demonstrate passive clustering and active switching of multiple μPPT
propellant/electrode assemblies with a single PPU and discharge capacitor. These options
promise to greatly increase the performance and flexibility of the μPPT, at only a modest cost in
complexity. It should be noted that the baseline μPPT is a very simple device, assembled from a
small number of commercial, off-the-shelf components.
Finally, a flight μPPT was developed for the US Air Force Academy’s FalconSat 3 spacecraft.
Construction and testing of this thruster were done by the Busek corporation, pursuant to a license
from WER and this author. The actual flight unit was delivered on February 28, 2006, has been
integrated with the spacecraft, and is expected to fly in early 2007.

144
Bibliography

[1] Antonsen, E.L., Burton, R.L, et al, “Effects of Post-Pulse Surface Temperature on Micro-
Pulsed Plasma Thruster Operation”, AIAA 2004-3462, 40
th
AIAA Joint Propulsion
Conference, Ft. Lauderdale, FL, July 2004
[2] Antonsen, E.L., Burton, R.L., et al, “Herriott Cell Interferometer for Unsteady Density
Measurements in Small Length Scale Thruster Plasmas,” , AIAA 2000–3431, 36
th
AIAA
Joint Propulsion ConferenceHuntsville, AL, July 2000.
[3] Antonsen, E. L., “Herriott Cell Interferometry for Pulsed Plasma Density Measurements,” MS
Thesis, University of Illinois at Urbana–Champaign, 2001.
[4] Boyd, I. D Far Field,” Journal of Spacecraft and Rockets, Vol. 37, No. 3, 2000, pp. 399–407.,
Keidar, M., and McKeon, W., “Modeling of a Pulsed Plasma Thruster from Plasma
Generation to Plume
[5] Blandino, J.J. and Cassaday, R.J., “Propulsion Requirements and Options for the New
Millenium Interferometer (DS-3) Mission”, AIAA-1998-3331, AIAA Joint Propulsion
Conference, Cleveland, OH, July 1998
[6] Burton, R.L., and Bushman, S.S., “Probe Measurements in a Coaxial Gasdynamic PPT”,
AIAA-99-2288, 35
th
AIAA Joint Propulsion Conference, Los Angeles, CA, June 1999
[7] Burton, R.L., and Turchi, P.J., “Pulsed Plasma Thruster,” Journal of Propulsion and Power,
Vol. 14, No. 5, Sept. 1998, pp. 716–735.
[8] Bushman, S.S., and Burton, R.L., “Arc Measurements and Performace Characteristics of a
Coaxial Pulsed Plasma Thruster”, AIAA 98-3660, 34
th
AIAA Joint Propulsion Conference,
Cleveland, OH, July 1998
[9] Caceres, M, “A Starting Point for Satellite Market Forecasts”, Aerospace America, October
225, pp.11-13
[10] Herriott, D.R., Kogelnik, H., et al., “Off Axis Paths in Spherical Mirror Interferometers,”
Applied Optics, 3, 1964, 523
[11] Keidar, M. and Boyd, I.D., “Device and Plume Model of an Electrothermal Pulsed Plasma
Thruster,” AIAA 2000–3430, 36
th
AIAA Joint Propulsion Conference, Huntsville, AL, July
2000
[12] Keidar, M. and Boyd, I.D., “Electromagnetic Effects in the Near Field Plume Exhaust of a
Pulsed Plasma Thruster,” , AIAA 2001–3638, 37
th
AIAA Joint Propulsion Conference, Salt
Lake City, UT, July 2001.
[13] Keidar, M., Fan, J., Boyd, I.D., and Beilis, I. I., “Vaporization of Heated Materials Into
Discharge Plasmas,” Journal of Applied Physics, Vol 89, Issue 6, 2001, pp. 3095–3098.
145
[14] Mueller, Juergen, “Thruster Options for Microspacecraft: A Review and Evaluation of State-
of-the-Art and Emerging Technologies”, AIAA 97-3058 33
rd
AIAA Joint Propulsion
Conference, Seattle, WA, June 1997.
[15] Schilling, J.H., Spanjers, G.G., and White, D; “Methods to Increase Propellant Throughput in
a Micro Pulsed Plasma Thruster”, United States Patent 6,769,241 B2, August 2004
[16] Schilling. J.H., Spores, R. A., and Spanjers, G.G., “Micropropulsion Options for the TechSat
21 Space-Based Radar Flight”, Micropropulsion for Small Spacecraft, AIAA Press, 2000 pp.
3-22
[17] Solbes, A and Vondra, R.J., “Performance Study of a Solid Pulsed-Fuel Electrothruster”,
Journal of Spacecraft, Vol 10, p. 406 (1973)
[18] Spanjers, G.G, McFall, K.A., et al, “Investigation of Propellant Inefficiencies in a Pulsed
Plasma Thruster”, AIAA-96-2723, 32
nd
AIAA Joint Propulsion Conference, Lake Buena
Vista, FL, July 1996
[19] Spanjers, G.G., Schilling, J.H, et al, “The Micro Pulsed Plasma Thruster”, Technical Report,
AD-A407730, AFRL Space and Missile Propulsion Division, June 1999
[20] Spanjers, G.G. and Spores, R.A, “PPT Research at AFRL,” 33
rd
Joint Propulsion Conference,
AIAA Paper 98–3659, Cleveland, OH, July 1998.
[21] Spanjers, G.G, et al, “Propellant Inefficiency Resulting from Particulate Emission in a Pulsed
Plasma Thruster”, Journal of Propulsion and Power, Vol. 14 no. 3, May-June 1998
[22] Vondra, R.J. and Thomassen, K.I., “Flight Qualified Pulsed Electric Thruster for Satellite
Control”, Journal of Spacecraft, Vol. 11 No. 9, September 1974, pp. 613-617
[23] Wertz, J. R., and Larson, W. J. “Mass Distribution for Selected Satellites”, Space Mission
Analysis & Design 3
rd
Edition, Microcosm Press, 1999, pp.894-896
[24]White, D., et al., “AFRL µPPT Development for Small Satellite Propulsion,” AIAA
2002- 2120, 33
rd
AIAA Plasmadynamics and Lasers Conference, Maui, HI, May 2002.

Abstract (if available)

Abstract In this dissertation, the development of the two-stage micro pulsed plasma thruster (uPPT) is described. This development followed from a research effort aimed at overcoming crippling performance limitations of the conventional PPT. Several fundamental causes of this performance shortfall were identified, and attempts to overcome them were uniformly unsuccessful. Furthermore, a trade study of other micropropulsion systems suggested that even an improved conventional PPT would find little use.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

A study of ignition effects on thruster performance of a multi-electrode capillary discharge using visible emission spectroscopy diagnostics

PDF

The cathode plasma simulation

PDF

Topics in high voltage pulsed power plasma devices and applications

PDF

Theoretical and experimental investigtion into high current hollow cathode arc attachment

PDF

The physics and application of compact pulsed power to transient plasma ignition

PDF

Experimental and numerical investigations of charging interactions of a dusty surface in space plasma

PDF

An experimental investigation of cathode erosion in high current magnetoplasmadynamic arc discharges

PDF

Experimental and analytical investigation of a ring cusp ion thruster: Discharge chamber physics and performance

PDF

Development of impulse measurement techniques for the investigation of transient forces due to laser-induced ablation

PDF

A theoretical and experimental investigation of cathode processes in electric thrusters

PDF

Particle simulations of colloid thruster beam

PDF

Laboratory investigations of the near surface plasma field and charging at the lunar terminator

PDF

Numerical and experimental investigations of ionic electrospray thruster plume

PDF

Mitigation of ion motion in future plasma Wakefield accelerators

PDF

Kinetic studies of collisionless mesothermal plasma flow dynamics

PDF

An investigation into magnetically induced extractor-less electrospray propulsion devices

PDF

Design and implementation of full solid state high voltage nanosecond pulse generators

PDF

Coupled multiple micro-resonators design and active semiconductor micro-resonator fabrication

PDF

Studies of nanosecond pulsed power for modifications of biomaterials and nanomaterials (SWCNT)

PDF

Designing an optimal software intensive system acquisition: a game theoretic approach

Asset Metadata

Creator Schilling, John (author)

Core Title Development of the two-stage micro pulsed plasma thruster

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Astronautical Engineering

Publication Date 02/14/2007

Defense Date 11/14/2006

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag micro PPT,OAI-PMH Harvest,pulsed plasma thruster

Language English

Advisor Erwin, Daniel A. (committee chair), Gundersen, Martin A. (committee member), Katsouleas, Thomas (committee member), Kunc, Joseph A. (committee member), Muntz, Eric Phillip (committee member)

Creator Email schillin@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m249

Unique identifier UC1148144

Identifier etd-Schilling-20070214 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-162270 (legacy record id),usctheses-m249 (legacy record id)

Legacy Identifier etd-Schilling-20070214.pdf

Dmrecord 162270

Document Type Dissertation

Rights Schilling, John

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu