Content uploaded by Gary R. Swenson
Author content
All content in this area was uploaded by Gary R. Swenson on Aug 27, 2014
Content may be subject to copyright.
Approved for public re lease; distribution is unlim ite d.
This effort was performed under contract NNX09CB37C with NASA Marshall Space Flight Center.
INITIAL DEVELOPMENT OF THE CUBESAIL/ULTRASAIL SPACECRAFT
R. L. Burton, J. K. Laystrom-Woodard, G. F. Benavides, and D. L. Carroll
CU Aerospace, L.L.C.
Champaign, IL
V. L. Coverstone, G. R. Swenson, A. Pukniel, A. Ghosh, and A. D. Moctezuma
University of Illinois at Urbana-Champaign
Urbana, IL
ABSTRACT
The UltraSail spacecraft can potentially achieve square-kilometer sail areas by deploying sail blades
between satellites separated by as much as 5000 m. In the UltraSail version, the spacecraft spins about
a central hub to flatten the blades against photon pressure, supported by tip-satellites. The sail is
launched in a rolled-up “window-shade” configuration, allowing sail storage and deployment to very large
areas without folding. Spacecraft control of the thrust vector by spin-axis slewing is achieved by twisting
the blades with the tip-satellites. Propellant mass fraction is very small, allowing use of cold gas even for
high-ΔV interplanetary missions. CubeSail is designed as a low-cost demonstration of UltraSail, using
two near-identical CubeSat satellites to deploy a 250 m-long, 20 m2 reflecting film between them. A
minimum altitude of 750 km is needed to minimize aerodynamic forces. The two satellites are launched
as a unit, detumbled, and separated, with the film unwinding symmetrically from motorized reels in the
payload bays. It is found that gravity-gradient is preferred over spinning as a way to provide the
necessary tension in the film. CubeSail design and experimental development is described.
INTRODUCTION
Many solar sail designs [Vulpeti, 2008] employ masts and rigging for sail deployment. Mast systems
limit sail performance in two ways, in that the mass of the mast reduces spacecraft acceleration, and the
mast length is limited structurally, therefore limiting the sail area.
Attempts to overcome mast limitations go back to the Heliogyro [MacNeal, 1967][MacNeal, 1971]
(Fig. 1), which employed radial blades attached to a central core, centrifugally spun to provide stiffness.
While the Heliogyro could in principal achieve very large areas (MacNeal envisioned blade lengths up to
30 km), blade control issues existed. Achieving the initial spin-up would be difficult, and the damping of
radial waves is difficult at the free end. Blade response is relatively slow, requiring approximately 2/3 of a
revolution (~2 radians) for a commanded pitch wave to travel from the core to a blade tip and back.
UltraSail (Fig. 2), which builds on Heliogyro heritage by using a spinning blade system attached to a
central hub, was conceived to improve control by adding a low-mass controllable satellite (tipsat) at the tip
of each blade [Botter, 2008][Hargens-Rysanek, 2007]. By initiating blade control at the tip, instead of at
the hub, blade response is improved to approximately 1/3 of a revolution (~1 radian) for a commanded
pitch wave to travel from to a blade tip to the hub. This added control reduces the spin rate, enables
blade pitch control for spin-up.
2
Figure 1: Heliogyro.
Figure 2: UltraSail launch configuration.
ANALYSIS IN TERMS OF THE ROCKET EQUATION
The spacecraft designer can consider an electric propulsion system on a high ΔV spacecraft in terms
of three subsystems:
total mass = transfer mass (bus + payload + power source) + propellant
For a solar sail vehicle:
total mass = transfer mass (bus + payload) + solar sail
The solar sail thus plays the same role as propellant on a conventional spacecraft. A major advantage of
the “propellantless” solar sail is that the need for a relatively massive and expensive power source is
avoided.
The equivalent specific impulse and power of a solar sail are found from the rocket equation, written
as:
characteristic velocity (ΔV) = (g · specific impulse) ln[(total mass)/(total mass – solar sail mass)]
so that the effective “exhaust velocity” Ue = gIsp is:
Ue=gIsp =!V / ln mo
mo"mss
#
$
%&
'
(
By analogy with electric systems, in terms of thrust T the solar sail power is
P
ss =1
2TUe
. Writing thrust in
terms of the solar pressure Po as
T=2P
o
A
ss
=2P
o
m
ss
/!
3
where σ is the areal specific mass in kg/m2. For mss/mo = µss, the power is:
P
ss =1
2TUe=P
omss
!V
"ln 1
1#µss
$
%
&
&
'
(
)
)
For systems with µss << 1:
P
ss !
P
omss"V
#µss
To compare solar sail systems with electric systems, it is convenient to compare the specific mass α
[kg/W]:
!=
m
ss
P
ss
=
"µ
ss
P
o
#V
As an example, consider a sail with an areal density of 5 [g/m2], µss = 0.1, P = 4x10-6 [N/m2], and ΔV = 25
km/s, giving α = 5 g/W = 5 kg/kW.
An α of 5 kg/kW is very competitive when compared to electric systems, for which α is typically 20
kg/kW [Bonfiglio, 2005]. For example, a mass comparison of (a ΔV) = 20 km/s solar electric and solar sail
mission, where the power level is 30 kW, is given in Table 1.
Table 1: Mass comparison of solar electric vs. solar sail missions.
NEXT Solar Electric (SEP)
Solar Sail (SS)
Mission ΔV [km/s]
20
20
Bus + payload mass [kg]
1600
1600
Launch mass [kg]
3300
1800
Acceleration [mm/s2]
0.21
0.21
Thrust at 1 AU [N]
0.70
0.35
Propellant + tank [kg]
1300
0
Propulsion mass [kg]
400
200
Eff. Specific impulse [s]
4000
20,000
Effective power [kW]
20
35
Sail area [m2]
--
40,000
Launch Vehicle
Delta IV Medium
Falcon 9
Launch Cost
$160 M
$50 M
Propulsion Cost
(est.) $100 M
TBD
ULTRASAIL CONCEPT
UltraSail was conceived as a way to achieve extremely high solar sail performance by minimizing sail
support hardware coupled with 1 km2-class sail areas [Burton, 2005]. Sail material is mounted on
multiple reels, each with a width of 5 – 10 m, and deployed to a blade length up to 5000 m. Deployment
4
and blade control is enabled with a satellite (tipsat) attached to each blade tip (Figs. 3 – 4). Calculated
performance exceeds that of solar electric propulsion systems.
Figure 3: UltraSail deployment.
Figure 4: UltraSail blades deploy from a central hub to
achieve a km2-class sail area.
The baseline design for UltraSail assumes the use of a coated polyimide film with ripstop. UltraSail
blade stiffness is achieved by spinning the spacecraft. Total film stress due to rotation is a few per cent of
yield, not including the strength added by the ripstop.
Fully deployed, each tipsat performs a circular orbit about the central hub, with rotational axis
generally pointing at the sun. The tip-satellite, a lightweight beam truss with onboard propulsion, performs
the following functions:
1. A stable, stiff attachment point for the film end.
2. Bus for propellant, solar panels, thrusters, and a metrology system.
3. On-board thrust to initiate film roll-out and initial blade spin.
4. Satellite metrology system to provide blade tip position, velocity and acceleration.
5. Twists (pitches) the blades to induce torques for spin-stabilization, using pitch thrusters.
6. Satellite centrifugal force flattens blade to increase photon thrust.
7. Provides continuous plane-change control for Sun-orientation of spacecraft spin axis.
8. Blade camber control to stabilize orientation along the blade axis.
Therefore, if the tip-satellite is very small, the possibility exists for extremely low areal densities
approaching the areal density of the film itself.
The rotation of the film and blade mass provide centrifugal force that tends to prevent the blade from
blowing back due to the solar pressure Po. Dynamic models have shown that the optimum spin rate
achieves a total centrifugal force on a blade of 3 – 5 times the solar pressure force. For km-long blades
this results in a rotational period of 1 – 2 hours, and a tip speed of 10s of m/s.
SOLAR RADIATION PRESSURE
A right-handed orthogonal coordinate system was constructed by defining the z-axis to coincide with
the spin axis of the UltraSail. Positive z was defined to be in the direction away from the sun. The x-axis
was defined to be normal to the z-axis in the direction of an undeflected blade. Note that this coordinate
system rotates with the blade. Because the rotating blades are not rigidly attached to the hub, they will
be deflected away from the Sun due to the photon force until equilibrium is achieved between centrifugal
5
force and photon force. In this analysis, the blades were assumed to remain straight, even though in
reality they will assume a catenary shape of the form:
z(x) =!P
o
"
2
cx
m
arctanh x #c 2m
sat
R+#cR
2
( )
( )
1/2
$
%
&'
(
)
#cR 2 m
sat
+#cR 2
( )
$
%'
(
1/2
+
ln "
2
m
sat
R+#cR
2
2! #cx
2
2
( )
$
%'
(
#
$
%
&
&
&
&
'
(
)
)
)
)
where xm = R[1 – msat/(M/N + msat)]. The above equation gives deflection in the z direction as a function
of the x coordinate of the point along the centerline of the sail. For small deflection angles, the pressure
on a catenary blade shape versus a straight blade will be nearly identical. The necessity to display
maximum possible solar sail area to the Sun precludes large deflection angles, and thus the straight
blade assumption is valid (Fig. 5).
0
200
400
600
800
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5 8 8.5 9 9.5 10
x, km
z(x), meters
w/o Blade Mass
w/ Blade Mass
Figure 5: Catenary Shape. The tipsat deflection angle (w/ blade mass) is 2.6 deg.
In the UltraSail design, the amount of this deflection can be controlled by the angular velocity of the
blades (at higher angular velocities, the deflection angle, ζ, will decrease due to higher centrifugal force).
As will be discussed later, part of this work involved finding the amount of centrifugal force that yielded
the optimal deflection angle, based on increased or decreased fuel consumption of the tip-satellites and
the associated changes in tip-satellite mass.
Subsystem mass analysis for the tip-sat and hub was conducted to provide baseline masses for the
calculations performed above. Subsystems studied for the tip-satellites include propulsion, avionics,
power, structure, and the blade deployment mechanism/compliant attachment.
The main structure of the tip-satellites consists of a simple truss, constructed of either aluminum or
carbon fiber tubes. The structure is 5 m long with a 0.5 x 0.5 m2 cross section. The truss structure
consists of 10 half meter cube elements, with cross members on each face of the cube, Fig. 6. Using half
inch carbon fiber tube with a 0.8 mm wall thickness for the truss, yields a mass of 4.6 kg. Fittings,
attachment points, and passive thermal control systems bring the total structure mass up to 8.0 kg.
Figure 6: Tipsat. 27 kg tipsat supports 71 kg of film.
6
The avionics system is relatively complex because of the need for relatively decent positioning data.
Therefore, each tip-satellite has two Sun-sensors and two star-trackers, for redundancy. Further, the
positioning information is gathered with a microwave ranging system as described in the ranging section.
Computation is done at the hub in near-real time to achieve nearly autonomous operation; a modest
communications system is necessary to relay the information between hub and tip-satellite. These
communications are also necessary for the relay of other pertinent subsystem data, commands, and state
of health.
The power subsystem consists of body-mounted solar panels and associated power processing
equipment. Beginning-of-life estimates for solar panel specific power are 50 W/kg. Power requirements
for the tip-sat are assumed to be 50 W, yielding 1 kg in solar arrays per satellite. Additional processing
equipment and batteries are another 2 kg.
The propulsion system on the tip-sat consists of 2 thruster units mounted on either end of the truss
structure, providing thrust on the order of 0.1 N. These units utilize catalyzed nitrous oxide (N2O), to yield
approximately 150 s Isp. The attitude control system consists of nearly identical thrusters, also mounted
on the ends of the truss and providing thrust on the order of 10 mN, using cold butane. Depending on
film thickness, mission lengths and force ratios, the propellant mass varies between 1 and 20 kg.
Assuming a pressure fed system, with a chamber pressure of 750 psi and Titanium tanks tank mass was
determined as a function of initial propellant mass, and ranged from 1 – 6 kg.
It was shown that a tip-satellite dry mass of approximately 42 kg and a hub dry mass of just under 50
kg is possible with state of the art technology. This is encouraging from the standpoint of minimizing areal
density of the combined blade and tip-sat system.
The final part of this analysis was to determine the most appropriate force ratio for a given mission
length. The force due to solar pressure for the baseline UltraSail design when the deflection angle ζ is
zero is 0.228 N, the maximum possible force. For large deflection angles, (low force ratio) the blade
length needs to be increased as the force drops with the cosine of the deflection angle. For each new
blade length, and hence increased blade mass, the angular velocity of the system was assumed to
remain the same. Because blade length increased and angular velocity remained the same, the tip
velocity increases. This change requires a new value for the amount of propellant and tank mass. A new,
greater, tip-sat mass was then calculated and the problem repeated until it converged. These increases
were summed for all four satellites and blades.
The total mass of the four tip satellites and blades are plotted as a function of force ratio for varying
mission lengths in Fig. 7, which shows a force ratio for minimum tip-satellite and blade mass. This
optimum force ratio runs between 3 and 3.5 for the higher mission lengths. For the two year mission, the
minimum is at a force ratio around 5. However, the payload mass gain is so minimal between 3.5 and 5
that it does not seem necessary to operate the UltraSail at higher force ratios. For the baseline UltraSail
design, a force ratio of 3 was adopted.
7
400
420
440
460
480
500
520
540
560
580
01234567
Force Ratio
4!(mblade + msat ), kg
8 years
6 years
4 years
2 years
Figure 7: Mass of all Tip-Satellites and Blades vs. Force Ratio for 4 Mission Durations.
VACUUM DEPLOYMENT EXPERIMENT
One of the most challenging aspects of the UltraSail project is the deployment of a potentially
kilometers-long sail from a rolled up configuration. The Vacuum Deployment Experiment (VDE)
measures the required force to unwind the sail material from the reel in a vacuum environment as a
function of speed. Also, the experiment is designed to be portable enough to allow it to be placed on
board a micro-gravity simulation aircraft, to study unrolling in a zero-g environment.
Two reels are placed in a vacuum chamber (Fig. 8). One reel is pre-wrapped with 100 meters of 30
cm wide, 2.5 micron film from SRS technologies, and the other reel acts as a take up. The reels are
synchronized by a timing belt, and connected to a precision variable speed motor on the outside of the
tank by a ferro-fluidic rotary motion vacuum feedthrough. The force transducer shown in Fig. 8 indirectly
measures the tension in the sail material at different unwinding speeds, to a maximum force of 125 g. By
closely measuring the angles of the film between the reels and the roller, along with the measured force,
the film tension can be calculated, which is equivalent to the peel force. The unwinding force overcomes
the peeling force, which is the sum of the electrostatic and adhesive forces between the sail surfaces.
The tension measurement system is shown in Fig. 8; Fig. 9 shows the vacuum chamber.
Deployment
Reel
Take-up Reel
!
Force
Gauge
Pivot
Idler
Roller
Deployment
Reel
Take-up Reel
!
Force
Gauge
Pivot
Idler
Roller
Figure 8: Tension measurement system.
8
Figure 9: Vacuum Deployment Experiment (front and back).
The peeling force is in general an unknown function of unwinding speed. The required unwinding
force will be provided by tip satellites and/or by centrifugal force. The expected tension range is very
small (0 - 10 mN), so the experiment was designed to measure these small forces. Unwinding speeds up
to 1 m/s were simulated by the experiment.
Initial experiments were performed with a filled teflon slit material. It was found that the peel force
increased with time, and soon became excessive, due to electrostatic forces. Efforts to drain off
electrostatic charge with a grounding brush failed, and the slit was then fabricated with aluminum.
The aluminum slit was highly successful. The peel force was too small to measure, and no
electrostatic charge buildup was observed. It was concluded that the film could be deployed even with
the small forces available in orbit due to gravity gradient or spacecraft spin.
CUBESAIL CONCEPT
It was concluded that a low cost way to test the UltraSail concept would be to use two CubeSats,
nearly identical, with the film deployed between them (Fig. 10). The spacecraft is called CubeSail. Each
CubeSat holds approximately 8 cm x 125 m (10 m2) of 6.2 mm double-coated Mylar film. The spacecraft
mass is ~2.5 kg, and the film mass is ~90 g. The total mass of the film, deployment equipment, and
associated electronics is ~500 g.
9
Figure 10: Artist’s rendition of CubeSail solar sail deployment.
The relatively slow emergence of solar sailing as a viable space propulsion method can be traced to
three challenges: low technology readiness level, complications related to stowage, deployment, and
support of large flexible structures, and control of the large flexible spacecraft. To address these
challenges, we proposed a small-scale (20 m2) demonstration of deployment, sail performance during
nominal operations, and orbital maneuvering in low Earth orbit (LEO). The low Earth orbit environment
poses unique challenges for a 20 m2 solar sailing spacecraft, including time-varying lighting conditions,
residual atmospheric drag disturbance torques, and potential of film damage due to micro-meteorite
debris. Despite these challenges, deployment into a low Earth orbit is chosen in order to take advantage
of highly reduced launch costs as secondary payload, and to advance the technology readiness level
(TRL) of several CubeSail subsystems. The spacecraft fully conforms to the CubeSat Design
Specifications [Cal Poly SLO, 2009] and can thus be integrated with the space-qualified Poly-Pico
Satellite Deployer (P-POD), shown in Fig. 11. This ability not only reduces complexity in integrating our
spacecraft with the launch vehicle, but also instills confidence in the safety of the primary payload and
facilitates unproblematic deployment into a desired orbit.
Figure 11: Poly-Pico Satellite Deployer [Cal Poly SLO, 2010].
The CubeSail mission is a first in a series of increasingly complex demonstrations leading up to a full-
scale UltraSail mission. The primary mission objective is to test the reel-based stowage and deployment
mechanism that eliminates traditional sail support structures such as booms, masts, stays, or guy-wires.
10
The film is wound onto two motor-driven, variable-speed reels, each placed in a 15x10x10 cm tip satellite.
Internal to the reel is the motor, magnetic encoder for measuring deployment rate and length, and related
wiring. The sail material is a polyimide-based polymer coated on the front side with aluminum to achieve
maximum reflectivity and to ensure that equilibrium temperature is within the operating range of the base
polymer. In addition, the film is reinforced with ripstop thread to prevent tear propagation in the event of
film damage. The resulting total film thickness is 6.2 µm. The payload bay constraint on film width of 78
mm results in total film length of 260 m, split evenly between the two CubeSats.
The top-level mission sequence of the CubeSail begins with P-POD ejection and separation of the
spacecraft from the upper stage of the launch vehicle, followed by bus power-up and internal health
checkout procedures. The batteries are allowed to recharge and initial ground communication is
established using an omnidirectional antenna. After successful checkout and verification of nominal
operations of all subsystems, the satellite initiates despin and reorientation. During this phase, the two
CubeSats are held rigidly together using the SRU and are oriented with its long axis along the nadir
direction with zero rotational velocity. Attitude determination and control methods, including sensor suite
and actuators, are discussed in greater detail in a subsequent section.
Once CubeSail is despun and oriented along the nadir, the batteries are allowed to fully recharge and
the separation sequence is initiated. First, film tension is released. The SRU motor is then started and is
rotated four turns until the satellites are separated. Guide pins ensure both spacecraft are locked in the
pitch direction and that the separation occurs linearly. The film is now in tension from two separation
springs, which provide initial separation velocity.
After the SRU operation is complete, film deployment is initiated by engaging the reel motor and
running it at a prescribed time-varying rate until full 260 m of sail are deployed. The gravity gradient force
ensures that sufficient tension exists in the film to prevent excessive billowing due to solar radiation
pressure and residual aerodynamic drag. Details regarding the deployment dynamics are presented in a
subsequent section.
The desired insertion orbit is selected based on several design criteria, the most important of which is
a constant lighting condition. Avoiding transitions between sunlit and dark portions of the orbit minimizes
time-varying dynamic loading of the sail. In orbits which do not guarantee constant lighting conditions,
the sail billowing will increase during the sunlit portions and decrease during the dark portion, creating an
accordion-like effect on the CubeSats. The non-constant force ratio between solar radiation pressure and
gravity gradient can cause damage to the film and must be minimized. Additionally, orbits with constant
lighting avoid thermal cycling of the sail material, and offer much easier planning of orbital maneuvers.
As a result, CubeSail’s target orbit is the sun-synchronous terminator orbit, shown in Fig. 12.
Figure 12: CubeSail in sun-synchronous terminator orbit. Image is not to
scale and shows no billowing effects. The yellow line on the surface
indicates the solar terminator.
11
The desired insertion altitude into the sun-synchronous orbit is selected to achieve a baseline one
year orbital lifetime. The plot and associated table, Fig. 13 and Table 2, show the relationship between
insertion altitudes and orbital lifetime for various attitudes and times within the solar cycle. The variations
in solar flux over the 11 year sun cycle cause fluctuations in the Earth’s atmospheric density by as much
as two orders of magnitude and therefore must be carefully taken into account. The calculations were
performed using the Naval Research Laboratory’s NRLMSISE2000 atmospheric density model [Naval
Research Laboratory, 2000] and the Analytical Graphics Incorporated Satellite Tool Kit Lifetime analysis
tool. The shaded areas in the table represent altitudes at which the spacecraft will survive at least 1 year
before deorbiting, for each of the described attitudes. It is important to note that for the worst case
scenario when the sail is face-on to the velocity vector, the spacecraft must be deployed into a 800 km
orbit during solar minimum and 900 km orbit during solar maximum.
300 400 500 600 700 800 900
10
-4
10
-3
10
-2
10
-1
10
0
10
1
10
2
10
3
Starting altitude [k m]
L
i
f
e
t
i
m
e
[
y
r
s
]
Cd=2
Flat, specularly reflective sail
Solar Min Year: 2008
Solar Max Year: 2014
Density Model: NRLMSISE 2000
Edge-on Solar Max
Edge-on Solar Min
Face-on Solar Max
Face-on Solar Min
Figure 13: Orbital lifetime predictions for various
attitudes, solar cycle intervals, and altitudes.
Table 2: Orbital lifetime with respect to insertion
altitude and solar cycle.
Alt
(km)
Orbital Lifetime (yrs)
Edge-on
Sol Min
Edge-on
Sol Max
Face-on
Sol Min
Face-on
Sol Max
400
1.4
0.6
0.004
0.0012
500
6.7
4.9
0.034
0.0073
600
27.2
22.4
0.176
0.0036
700
106
92.5
0.508
0.137
800
317
316
1.169
0.386
900
>300
>300
2.044
1.108
Although the launch date within the solar cycle has a significant effect on the orbital lifetime, the sail
attitude with respect to the velocity direction has an even greater impact. From the above table, during a
solar maximum period, if the film is deployed with the sail normal perpendicular to the velocity direction
(edge-on) and must survive for at least one year, it must be deployed into a minimum initial altitude of 500
km. In contrast, if the sail is deployed in the face-on configuration during the same period, it must be
inserted into a 900 km initial orbit. As a result of this strong dependence of orbital lifetime on CubeSail
attitude, the spacecraft is nominally flown in the zero pitch (edge-on) configuration. During orbital
maneuvers, the CubeSats can be pitched in opposite directions, inducing a twist in the film and creating a
net force. Appropriate pitching maneuvers are thus used to produce changes in orbital inclination and
altitude and demonstrate CubeSail’s solar radiation pressure maneuvering capability.
BUS AND GROUND COMMUNICATIONS
CubeSat is based on the IlliniSat-2 bus, designed to be a fully modular, highly adaptable system that
is fully compatible with the CubeSat and P-POD standards. The satellite bus consists of power,
command and data handling (C&DH), and attitude determination and control (ADCS) subsystems. The
IlliniSat-2 bus in a 1.5U structure has a mass of about 900 grams and the bus electronics fit into less than
a 1U cube as shown in Fig. 14. As such, the bus provides excellent mass margin (~1100 grams) and
volume margin (~9x9x8 cm) for payloads/instruments. The system is constructed as a set of sub-
assemblies, allowing pre-integration testing of the payload and service modules independently. The
IlliniSat-2 bus builds on experience gained in the design of ION1 and makes extensive use of emerging
technologies that will serve to push the envelope of future CubeSat missions. Extensive ground testing of
the bus outlined in the supplementary document will ensure mission success, despite the lack of flight
heritage for the current design.
12
Figure 14: Generic IlliniSat-2 bus in a 1.5-U
configuration.
POWER SUBSYSTEM
The power subsystem consists of four body-mounted solar panels, a lithium-ion battery pack, and
supporting circuitry required to operate the spacecraft and associated science payload. Each solar panel
consists of 3 Spectrolab triple junction solar cells (25% QE) mounted on a light- weight carbon fiber panel.
The carbon fiber backing provides a significant mass reduction over typical FR4 fiberglass panels.
Maximum operating efficiency of each string of solar cells is maintained via power point tracking. Power
collected from the cells is directed through a battery charger to maintain and charge a lithium-ion battery
pack (7.4 V, 2200 mAh). Solar panels occupy the four largest of the six faces of the CubeSat. A six-
month simulation of a nominal mid-latitude orbit (45o inclination) indicates that 2 W of on-orbit average
power will be produced in the 1.5-U configuration, which is sufficient for the proposed CubeSail mission.
The power board has high efficiency (>85%) switching voltage regulators to provide 5 VDC for the
satellite bus. Additionally, 8 switched lines from the battery (7.4 VDC) are available for the radio and
additional heaters/high current loads as required. Each regulator has been designed for maximum power
efficiency and the power subsystem is intelligently controlled and monitored via a microcontroller on the
power PCB that coordinates power control with the C&DH motherboard.
C&DH SUBSYSTEM
The C&DH subsystem controls the operation of the spacecraft and provides communication with the
ground station. The C&DH motherboard includes a Texas Instruments OMAP5912 processor and system
memory, including random access memory (64 MB) for program operation and non-volatile flash memory
(> 1 GB) for data and program storage. Included in the C&DH subsystem are PIC microcontrollers that
reside on the power and ADCS PCB’s, which intelligently control those subsystems and communicate via
the board backplane with the main OMAP processor via the I2C protocol, a standard two-wire serial
protocol.
The OMAP5912 is a low power, high performance dual-core processor with an ARM9 32-bit
microprocessor and TMS320C55x digital signal processor (DSP) in the same chip. The ARM9 is the
central processing unit that hosts the Linux operating system, built on the 2.6 kernel, and associated
device drivers. The flight software, developed at U of I, coordinates internal and external data operations,
monitors spacecraft health, and collects, formats, and interfaces with the CubeSail control board, sending
commands and storing data (photos). The DSP performs the required calculations for the attitude
determination and control algorithms and serves as the terminal node controller/modem during satellite-
earth communications, providing a data rate of 2400 baud, using a FSK modulation scheme. The
capable DSP core in the OMAP5912 should allow other bandwidth efficient modulation schemes such as
QPSK and QAM to be designed and implemented as required.
The communication system is based on a half-duplex amateur band radio (Dataradio DM-3475) tuned
to ~437 MHz (the actual frequency is applied for though the FCC, coordinating with the ARC) with
13
variable 0.75-2 W transmit power. The radio is connected to the software TNC implemented in the DSP
and the AX-25 protocol is used for uplink and downlink. A circularly-polarized, crossed dipole antenna
has been constructed from a memory metal that is folded up against the satellite and is deployed using
nichrome wire and monofilament. The antenna was designed to have a good omnidirectional pattern to
ensure communication with the spacecraft regardless of the current attitude as shown in Fig. 15.
Figure 15: Beam pattern of crossed dipole antenna for
uplink and downlink communication on TIWS,
demonstrating its measured omnidirectionality.
ADCS SUBSYSTEM
IlliniSat-2 has full closed-loop three-axis attitude determination and control, implemented on board the
spacecraft via a magnetic control system. An integrated high-precision low power 3-axis magnetometer
(Honeywell HMC6343) and a coarse sun sensor are used for attitude determination. The information
from the attitude sensors is fed into a Kalman filter (KF) for attitude estimation, yielding an attitude
knowledge to within one degree per axis. The attitude estimate is passed through a linear quadratic
regulator (LQR) to determine the optimal control input for the magnetic torque coils. Both the KF and
LQR algorithms are to be implemented on the DSP. Magnetic torques are provided by torque coils
fabricated on a flexible printed circuit made from a polyimide film (Kapton) that is integrated into the solar
panels as shown in Fig. 16. The coils provide three-axes control, with redundant systems in the X and Y
direction. In addition to providing the torque coil, the flexible circuits also provide power routing from the
solar cells to the power board. Following P-POD deployment, the system will autonomously begin a
detumbling procedure to reduce rotation rates to 0.1 deg/s and prepare the satellite for science and
communication operations. After detumbling, the ADCS subsystem will maintain the attitude of the
spacecraft to within five degrees to comply with the mission requirements.
14
Figure 16: Torque coil implemented on flexible printed circuit.
STRUCTURE AND THERMAL CONTROL SUBSYSTEM
The spacecraft structure uses a system of rails and plates. The rails connect the sections together
and dictate the form factor of the satellite. The payload, separated from the bus by a middle plate,
interfaces directly with the top plate and the satellite bus interfaces directly with the bottom plate.
Figure 14 demonstrates the solid model of the full satellite assembly. All the supporting system
boards screw directly into the bottom plate, while the battery attaches to the middle plate. The solar
panels are connected using six bolted joints, while their power harnesses attach directly to the power
board in the stack. The payload is attached to the top plate. Power and data connections for the payload
are provided via a wiring harness, accommodated by a channel in each of the bus PCB’s and the center
plate. The bottom plate serves as a radiator, so the selection of an appropriate optical coating can aid in
the necessary thermal control. Heat from the PCB’s and other components are conducted along
embedded thermal planes to the standoffs, which then conduct the heat directly to the bottom plate. Heat
from the batteries conducts through the middle plate, down the rails to the bottom plate, and heat from the
payload is conducted down the rails to the radiating bottom plate.
Initial analysis has shown that the structure meets all mechanical requirements, and has previously
passed a preliminary vibration test. The initial thermal models are promising, however additional
simulation are required before the system is ready for a thermal vacuum test.
MISSION OPERATIONS, GROUND STATIONS, DATA DISTRIBUTION
The IlliniSat-2 radio transceiver is a commercially available data radio operating in the 70 cm band.
An application will be filed with the FCC to obtain an amateur band “space station” frequency allocation
and a pre-space notification will be made to the FCC no later than 90 days before integration with the
launch vehicle according to 47 C.F.R. § 97.207 of the FCC rules. The amateur licensed frequency range
in the 70 cm band allowed for “space stations” according to 47 C.F.R. § 97.207 is 435-438 MHz, all of
which is accessible by our selected radio.
The satellite ground station was constructed several years ago in preparation for the launch of ION1.
Students and staff have had great success contacting on-orbit amateur band satellites, and verifying
proper station operation. The primary station radio is dual band 2 m/70 cm with enhanced satellite
functionality (Icom IC-910H). Two high gain Yagi-Uda antennas (one for each band) are mounted on a
15
motorized elevation/azimuth rotator atop a 35-foot antenna tower on top of a four-story building on the
Illinois campus. Commercially available satellite tracking software is used to determine overhead pass
times and control azimuth and elevation rotators for the ground station antennas tracking the satellite.
The software automatically calculates anticipated Doppler shift in the signal and tunes the radio
appropriately.
TECHNOLOGICAL READINESS AND HERITAGE
Small satellites and payloads in the 1-2 kg class, called CubeSats, and 20-30 kg, called nanosats,
have been under development at the University of Illinois since late 2001. The ION1 CubeSat was a 2U
CubeSat with a photometric remote sensing instrument. ION1 was lost in the failed DNEPR launch
attempt on July 26, 2006. The IlliniSat-2 development began in the fall of 2005, and primarily focused on
the development of a new spacecraft bus and supporting equipment. The University of Illinois has also
continued work on additional payloads, and partnered with Taylor University on other remote sensing
payloads for the TEST Nanosat. Faculty program direction is provided by G. Swenson and V. Coverstone
who oversee and provide guidance for the spacecraft bus development.
The past several years have been dedicated to the development of the IlliniSat-2 bus, which is near
completion. First revision PCB’s are nearly prepared for manufacture and the spacecraft structure is
mature, machined, and has completed a first round of vibration testing.
GROUND FACILITIES
The University of Illinois currently has two laboratories for CubeSat development and testing. A
hardware lab with 4 electronics workstations, soldering facilities, a Class 10,000 clean room, a thermal
vacuum chamber, and a magnetic test facility are dedicated to hardware design and testing. A software
development/modeling lab has an additional seven computer workstations equipped with advanced
engineering software packages, provided by the College of Engineering. General software tools available
include compilers and development environments, MATLAB, Pro/Engineer, Unigraphics, NX Space
Thermal, and EagleCAD for hardware and software simulation and verification. Should the need arise,
the project also has access to cluster computing to handle the larger simulations and analysis.
The Illinois CubeSat program has an operating ground station for tracking and communicating with
the 430-MHz frequencies. Many of our team members are active in the amateur radio community and
have the required licenses of operating frequencies needed for ION1 operations. Since a standard
amateur radio packet system is being used, the world-wide amateur radio community can be involved in
the monitoring of the spacecraft. In addition, the ground system is fully controllable via the Internet, for
those granted access privileges. This level of accessibility allows students the potential opportunity to
partner with student groups at other universities around the world for satellite operations, and increase
the daily window of opportunity for satellite communication.
PAYLOAD DESIGN, REEL AND SLIT
The payload design for CubeSail consists of three main components: the film reels, the separation
release unit (SRU), and the camera. The ree l assembly, Fig. 17, must fit within a 90 mm x 90 mm x
84 mm payloa d volu m e. The width of the bobbin with in the f langes is 81 mm and the film width is
77 mm to allow for 2.0 mm separation b etween the film and the flange.
16
Figure 17: Bobb in design.
As shown in Fig. 18, the bobbin assembly includes two open bearings and a motor assembly.
(a)
(b)
(c)
Figure 18: Bobbin hardware components – (a) separated, (b) assembled, (c) installed.
A wound bobbin is shown below in Fig. 19. Note that there is a gap between both flanges and the
film, Fig. 20, as during testing it was determined that film unwinding was impeded when the sail material
touched the flange/edge of the bobbin.
Sail film
wound on
bobbin
H-Frame supports
Bobbin
H-Frames
17
Figure 19: Aluminized Mylar® film wound on bobbin.
Figure 20: Clearance
between film and flange.
SEPARATION RELEASE UNIT (SRU)
The two CubeSats, A and B, are launched as a single unit, and must be separated in orbit after
ejection from the booster. This is accomplished with a Separation Release Unit (SRU), which is based on
release technology previously employed on CubeSats. The baseline design called for 2 SRUs, straddling
the film slit, with a mass budget of 50 g for 2 SRUs.
A mechanical simulation experiment called the Bearing Rail experiment (BRE) is being developed to
test the Separation Release Unit (SRU) (Figs. 21 – 24). Each CubeSat is suspended from a linear ball
bearing with the sail film extending between them, Fig. 21. The linear bearings slide along a 1-m
hardened steel rod. The far end of each rod is attached to a linear translation stage which allows
independent downward alignment of the far end of the rods such that friction can be canceled by gravity
and the satellites will separate at a constant rate. Calculations were performed to find the optimal
diameter of the rod such that there is less than 0.1 mm deflection caused by the weight of the cubesats
along the 1 m length; that diameter is 0.75". The satellites are attached to a yoke assembly via a pivot
bearing which allows pitching motion and then attached to the linear bearing via another pivot bearing
which allows yaw motion. Balance weights are used to bring the mass to ≈2 kg per satellite, with the c.g.
near the CubeSat geometric center.
Figure 21: Bearing rail experiment.
Figure 22: Bearing rail experiment mounted to lab wall.
Satellite A
Satellite B
SS Rod
Linear Bearing
Linear Translation
Stage
18
Figure 23: BRE, end view.
Figure 24: Bearing Rail Experiment.
The SRU is under development. A TiNi Aerospace shape memory alloy pinpuller device had been
evaluated, but it was decided that an SRU would have to be designed.
SATELLITE RELEASE MECHANISMS
The baseline SRU design called for the separation springs for the two spacecraft to be secured and
held in place with a monofilament, held in close contact with a short length piece of small-diameter
nichrome wire. When ready for deployment, current would be passed through the nichrome wire, heating
and melting the monofilament, releasing the springs. This method has been used in the CubeSat
community for antenna deployment for some time, and is so used on the Illinois CubeSat.
During launch and at booster ejection, the film, which is located between the two connected satellites,
is kept at a low-level of tension. The separation scenario is: a) the SRU wire is heated, breaking the
SRU; b) the reel motors are started at a predetermined rate, e.g. 1 cm/s; c) the springs provide a
separating force, and d) separation velocity is maintained by the speed of the motors. The SRU and its
spring are designed with respect to the center of gravity (CG) so as to minimize rotational torque on
CubeSats A and B. Once separated sufficiently, film tension is maintained either by spin- or gravity-
gradient stabilization.
19
A second approach was based on the tensile failure of heated Nitinol shape memory alloy, as shown
schematically in Fig. 25. This approach had high power consumption, and was difficult to assemble.
Figure 25: CubeSat Nitinol separation release unit.
A third approach to the separation release unit (SRU) design, consisted of a 12 V solenoid, mounting
hardware, latching pin, microprocessor for solenoid activation, and switching device, Figs. 26 and 27.
Prior to the hardware being installed in the satellites, the solenoid was bench-tested with the SRU
hardware and RC circuitry and deemed to be functional and adequate for initial BRE testing.
Figure 26: SRU installed.
Figure 27: BRE with reel assemblies installed;
balanced.
Three primary versions of the solenoid separation release unit (SRU) are shown in Fig. 28. The first
generation (Fig. 28a) had tolerances that were too large. The second generation (Fig. 28b) had tighter
tolerances, but resulted in too much friction between the pins which impeded retraction of the solenoid
pin. The third generation conical design (Fig. 28c) caused tipping of the two satellites resulting in friction
between the SRU pin and the sleeve bearing. Because of the problems associated with the solenoid
SRU, another SRU concept has been implemented, the motor SRU.
SRU latching pin (left)
& corresponding plate
hole (right)
Separation
springs
20
(a)
(b)
(c)
Figure 28: Design progression and separation steps of the SRU and solenoid pins.
The current generation SRU consists of a lead screw and gear motor (the same as used in the sail
film bobbins) on one payload plate and a threaded hole mounted to the other payload plate, Fig. 29. The
motor is remotely controlled to turn the lead screw pinned to its shaft, which then unscrews from the
threaded hole. As the lead screw turns out of the hole, the separation springs push the two satellites
apart.
Figure 29: Motor SRU concept.
The separation springs were chosen based on a desired separation speed of 5 – 10 cm/s. Figure 30
shows a drawing of the payload plate with holes for the separation springs. The locations were chosen to
be as close to the slit as possible to impart a minimum moment from unequal springs. The optimal
number of springs will be determined through experimentation on the BRE.
SRU gear motor
Lead screw pin
SRU lead screw
Plate w/
threaded
hole
Payload plate B
Payload plate A
Solenoid pin
SRU pin
1.
Solenoid
pin
retracts
2. Satellites separate
Sleeve
bearing
Scale: 0.5" ≈ 1/8"
21
Figure 30: Satellite payload plate with separation spring holes noted.
SRU CONTROL SYSTEM
The CubeSail SRU separation sequence uses two feedback sensors that accurately monitor the
position of both the screw and the bobbin motor. The sensors, together with a 16-bit microcontroller, a DC
motor controller, and H-bridge switches, make the following requirements of the control system possible:
1. The motor shaft of the screw motor does not stall when initially separating the satellites.
2. The screw motor rotates at 24 RPM and the bobbin motors in each CubeSat release the sail film
at 5 cm/s.
3. Unexpected frictional forces of the motor shaft with any other component of the satellites are
rejected. Unexpected frictional forces affecting the rotating shaft are compensated for by
increasing motor torque to keep velocity constant.
To avoid stalling of the gear motor when it starts unscrewing, the SRU is unscrewed with at least
three times the installed torque. The factor of three overcomes frictional forces when initially separating
the satellites. The target speed and position of the motors is met by using PID controllers, which
compensate for frictional forces that the rotating components might have against the satellite structure.
The separation sequence is commanded by the microcontroller, as follows:
1. The film is unwound by the amount slightly longer than the distance of the SRU lead screw stroke.
2. The gear motor starts separating the satellites until it completely disengages from the other
satellite, leaving the film in tension, produced by the separation springs.
3. The bobbin motors then separate the satellites at 10 cm/s from each other.
The microcontroller also accepts commands in an event where manual input is required.
The circuitry to control the SRU, called the Radio Controlled Open Loop Controller (RCOLC), is
shown in Figs. 31 – 33 and Table 3.
Separation
spring holes
22
Figure 31: RCOLC components.
Table 3: RCOLC components list.
#
Description
#
Description
1
SRU battery
8
Radio data harness
2
Motor harness for connecting motor with PCB
9
Radio receiver power harness
3
Motor for bobbin
10
Solenoid SRU active component
4
Radio control receiver
11
Solenoid connectors
5
9 V battery in-series connector for main power
12
Radio controller
6
9 V batteries for main power, 2x
13
Push buttons, 2x
7
Main power switch
14
PCB with open loop motor control radio
receiver, and SRU activation capabilities.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
23
Figure 32: Satellite A with PCB installed.
Figure 33: Close-up of Satellite A, PCB
installed.
CAMERA
As part of the payload, one camera on each satellite is used to observe the film as it unreels, as well
as the satellites as they separate from each other. Experiments were performed in the lab with the
objective of determining the picture clarity and visibility from the lens of the satellite camera. An ordinary
cell phone camera was used to help simulate the resolution qualities expected from the satellite camera,
since they share similar characteristics (Table 4 and Fig. 34).
Table 4: Camera comparison.
Cell Phone Camera
Satellite Camera
Manufacturer
Sony Ericsson
Electronics123.com
Model
w300i
C328R
Sensor
CMOS
CMOS
Resolution
VGA 640x480 pixels
VGA 640x480 pixels
Color
Color/Monochrome
Color/Monochrome
24
Figure 34: C328R Mini CMOS camera.
To simulate the twisting that may occur in space, pictures were taken at various angles rotated from
0° - 90°. This angle would be measured relative to the rotation of the two CubeSats. It was difficult to
keep the sail surface smooth, but the pictures were adequate for the purposes of the experiment (Figs. 35
– 36).
Figure 35: Rotated 0°.
Figure 36: Rotated 30°.
DETUMBLING AND RE-ORIENTATION OF THE CUBESAIL ASSEMBLY
The separation of the spacecraft from the upper stage of the launch vehicle will impart residual
torques onto the spacecraft, resulting in non-nominal attitude and potential three-axis tumble. Before the
sail deployment can begin, the combined CubeSail assembly must be de-spun and oriented with its long
axis along the local vertical direction. In practice, it is difficult to accurately predict the spacecraft’s initial
attitude and rotational rates apriori, requiring a robust controller that is capable of handling a wide range
of initial conditions.
Historically, the final stage and the P-POD for a CubeSat-class spacecraft impart at most 2.5 deg./sec
angular velocity to all three axes. Since initial despinning and stabilization of the CubeSail is critical to
mission success, the controller is designed with 100% margin and is capable of handling 5 deg./sec
rotational rates on all three axes.
CubeSail attitude control is accomplished with three-axis, variable-strength magnetic torquers. The
actuators consist of a novel design that imprints consecutive copper loops in a wind-down pattern on a
flexible circuit board, four layers deep, shown previously in Figure 16. The flexible torquers are mounted
behind the solar panels, include all necessary control circuitry, and are capable of producing maximum
magnetic dipole moments of 0.106 A-m2 (at imax = 0.4 A). Since both satellites have identical attitude
control capability and the torquers can be fired simultaneously, the initial detumbling and stabilization can
be accomplished using a maximum dipole moment of 0.212 A-m2.
The control problem is formulated using Linear Quadratic Regulator theory [Bryson, 1975][Brogan,
1991][Psiaki, 2001]. The desired performance of the system—in terms of power consumption and
stabilization time—is achieved by appropriate selection of the cost function and two weight matrices, Q
and R. In practice, selection of appropriate Q and R matrices (9x9 and 3x3 diagonal matrices
respectively) that yield robust results for a given orbit and a wide range of initial conditions is very difficult.
In order to allow easy adjustment to these matrices and evaluation of their performance, a Matlab-based
simulation that incorporates CubeSail system dynamics, models for the magnetic field, aerodynamic drag,
and gravity gradient, as well as LQR control theory has been developed. The simulation is used in
conjunction with a simple Genetic Algorithm (GA) [Goldberg, 1989][Pukniel, 2006] to find a near-optimal
set of Q and R matrices to minimize a despinning and orientation times, while not exceeding available
power limits.
25
The results presented below are sample simulation outputs based on the following assumptions. The
spacecraft was assumed to be a rigid body with the dimension of 30x10x10 cm and mass of 3 kg. At this
time, the center of mass was assumed to coincide perfectly with the geometrical center of the two-satellite
assembly. This is seen as an acceptable assumption since the CubeSat Design Specification requires
them to be within 20 mm of each other. Moreover, a parameter that represents any such deviation is
incorporated in the simulation code and is used in the aerodynamic torque calculations.
The spacecraft was assumed to be inserted into a typical CubeSat orbit, here assumed to be a 750
km altitude circular low Earth orbit inclined at 98°. The orbit was propagated for 15 hours and included
gravity gradient and aerodynamic drag disturbance torques, as well as the applied magnetic control
torque. The coefficient of drag was chosen to be 2.2 and the aerodynamic density was chosen to be the
mean of solar maximum and solar minimum densities at 750 km, or 4.5x10-14 kg/m3. The atmospheric
density values were calculated using the NRLMSISE2000 code.
In order to evaluate the robustness of the controller for various orbital deployment scenarios, the
simulation was run 1000 times with randomly generated initial attitudes. All simulations assumed the
worst-case scenario with initial rotational rate of 5 deg./sec on all three axes. Figures 37 and 38 are a
representative sample of the output from a single run.
Figure 37: Euler angle history (left) and angular body rates (right).
Figure 38: Duty cycle history (left) and torques on the spacecraft (right).
26
As can be seen in Fig. 37 (left), the spacecraft is stabilized in approximately 2 hours. The rotational
rates are reduced to under 1 deg./sec in 0.5 hours and further reduced to near zero in the next 1.5 hours
as seen in Fig. 37 (right). Figure 38 (left) shows that the controller initially used almost 100% of available
torque authority to reduce the rotational rates and then used finer control to achieve the nadir-pointing
attitude. The torques applied to the spacecraft, which include magnetic, aerodynamic, and gravity
gradient, are shown in Fig. 38 on the right. It is worth pointing out that prior to sail deployment, the drag
force is nearly negligible. This is due to the assumption that center of mass is aligned with the
geometrical center, causing the center of gravity and center of pressure to coincide and produce no
torque.
DEPLOYMENT ANALYSIS
Preliminary analysis of the CubeSail deployment methods yielded two viable options, shown in Fig.
39. In the spin-induced deployment method, shown on the left, the mated satellites are spun together
using onboard magnetic torquers and then released by engaging the SRU. The film is unwound until the
rotational rate of the system decreases to a prescribed threshold level that ensures the film does not
billow out excessively. At this time, the tip satellites rotate in opposite directions along the long axis of
each spacecraft, inducing a twist in the film. If properly oriented relative to the sun, the resultant pitch in
the blade forces the system to act as a ‘propeller’ and spin up. Once the desired rotational velocity is
achieved, the satellites return to their zero-pitch attitude and more film is unwound. This process is
repeated until the final spin rate and deployment length are achieved.
Figure 39: Sail Deployment Options: Spinning (left) and Gravity Gradient (right).
The centrifugal force of the rotating system provides the tension necessary to keep the film relatively
flat. The challenge of this approach resides in the complicated dynamics of two rotating CubeSats
connected by a non-rigid tether. Due to the limited responsiveness of the attitude control system, this
deployment method is reserved for a second CubeSail demonstration, after deployment hardware and
various other subsystems have been space qualified.
The second deployment method, shown on the right side of Fig. 39, utilizes gravity gradient force
between the two CubeSats to provide the necessary tension in the film. The mated satellites are oriented
into a nadir pointing attitude and separated using two compressed springs. The spring constants are
selected by modeling the dynamical behavior of the two satellites during deployment and imposing a final
boundary condition of zero relative velocity and zero displacement away from the local vertical. The initial
conditions assume the spacecraft are aligned along the local vertical resulting in one free variable, the
radial separation velocity, which is solved for numerically and then correlated to an appropriate spring
constant through the principle of conservation of energy.
The film is unwound using the reel motors until the entire sail length is deployed at which point the
gravity gradient force provides the desired tension. Fortunately, the rate of solar pressure force increases
at the same rate as the gravity gradient force and is equal to approximately five times the desired value.
There are several advantages to using the gravity gradient method, including: simplified deployment
dynamics (no spin-up maneuvers or complicated dynamics associated with rotating spacecraft), passive
attitude stabilization of the system, and overall mission risk reduction.
27
Preliminary analysis of the gravity gradient deployment dynamics of a 260 m film has been performed
for a spacecraft in an ecliptic orbit. The ecliptic orbit offered a relatively disturbance-free environment, in
which a sail flying with its edge to the velocity vector experienced minimal aerodynamic drag and minimal
solar radiation pressure forces. As such, it was initially selected as a candidate CubeSail orbit, but due to
non-constant lighting conditions has been eliminated in favor of the sun synchronous terminator orbit.
The non-linear equations of orbital motion were solved numerically to meet final boundary condition of
zero deviation away from nadir and zero relative velocity. The initial boundary condition assumed a
perfectly nadir oriented spacecraft and an initial separation velocity in the radial direction, the latter of
which is a free variable.
Figure 40 shows the results of the integration, where
l
and
l
!
represent the deployed sail length and
length rate and
θ
and
φ
represent in-plane and out-of-plane angles in the orbital frame. The deployment
is accomplished in approximately 1.3 hours and meets both boundary conditions with the initial separation
velocity of 0.175 m/s.
Figure 40: Gravity gradient deployment dynamics.
The gravity gradient deployment demonstrations is seen as an opportunity to test several of the
CubeSail subsystems, including magnetic torque actuation hardware and LQR performance, reel
assembly and motor, sail dynamical models, reaction wheel control (discussed shortly), camera
performance, and several bus subsystems. Validation of performance of these subsystems and any
consequent re-designs would pave the way for a more complex demonstration of the spin-induced
deployment and operations. This second demonstration would be a scaled-down version of the UltraSail
mission and would test control methods and performance of fully rotating system, a main attribute of the
UltraSail design.
ORBITAL MANEUVERS
The final phase of the CubeSail mission is the demonstration of propulsive capability using solar
radiation pressure. In order to achieve the desired thrust direction, the tip satellites are pitched in
opposite direction, inducing a twist in the film. The transfer maneuvers typically last several hours or days
(depending on available power), necessitating a fine control authority to maintain the desired pitch angle
in the presence of disturbing torques from aerodynamic drag and solar radiation pressure. Solving the
optimal control problem for magnetic torquers during orbital maneuvering is not only computationally and
power intensive, but also difficult to execute accurately due to dependence on a continuously varying
external magnetic field. As a result, pitching maneuvers necessary to maintain proper relative twist in the
film may have to be accomplished with miniature reaction wheels. Each CubeSail tip satellite would use
a single 10 mN-m-sec reaction wheel [Sinclair Interplanetary, 2009] (Fig. 41) to control the pitch angle.
28
The magnetic torquers would be used to dump momentum when the wheels become saturated or to
remove momentum bias.
Figure 41: Sinclair Interplanetary 10 mN-m-sec reaction wheel.
In order to predict environmental effects accurately on the CubeSail spacecraft from initial
deployment through nominal operations and orbital maneuvering, accurate force models have been
developed. The solar radiation pressure model includes non-ideal sail effects from reflection, absorption,
and re-radiation with the optical parameters listed in Table 5 [McInnes, 1999][Rowe, 1978][Wright,
1992][Wie. 2004].
Table 5: Optical coefficient of sample solar sails.
r
!
s
ε
f
ε
b
Bf
Bb
CubeSail
0.88
0.94
0.05
0.55
0.79
0.55
The notation in the above table is as follows:
r
!
is the coefficient of reflection, s is the fraction of
specularly reflected molecules,
ε
f and
ε
b are the front and back emissivity coefficients respectively, while
Bf and Bb are fractions of molecules that reflect in a semi-diffuse fashion from the front and back surfaces
respectively.
The aerodynamic drag model is based on the method of accommodation coefficients [NASA,
1971][Gaposchkin, 1994][Hughes, 1986] and takes into account interaction between the residual
atmospheric constituents and the sail surface. Factors that affect this interaction include angle of attack,
type and density of major atmospheric constituents at the desired altitude, orbital velocity, surface
temperature and level of thermal accommodation of impinging molecules, and fraction of molecules
reflected in the specular versus diffuse fashion. Sample calculations at an altitude of 800 km show that
depending on the angle of incidence,
α
AD, the equivalent coefficient of drag, CDEquivalent, varies between
2.85 and 1.35—a significant departure from the historically-used value of 2.2. CDEquivalent is computed by
equating aerodynamic drag forces computed using the accommodation coefficient method and classical
drag equation.
29
Figure 12: CDEquivalent obtained by matching aerodynamic forces computed using
accommodation coefficients and constant CD = 2.2.
The above models are used in conjunction with the Edelbaum low-thrust orbit transfer theory
[Chobotov, 2002][Edelbaum, 1961][Edelbaum, 1962] to compute transfer trajectories for three sample
cases. The cases are summarized in Table 6 and include: 1) 100 km altitude raise with minimal
inclination change (0.4º to maintain the sun-synchronous orbit), 2) 5º inclination change, and 3) 100 km
altitude increase combined with a 5º inclination change. The results indicate that even modest inclination
changes require long transfer times caused by minimal film area and hence available thrust. Time
histories of velocity, inclination, and altitude for the third orbital maneuver case are shown Fig. 43.
Table 6: Summary of results of 3 orbital maneuvers.
Scenario
i!
(º)
alt!
(km)
Transfer time
(days)
(1)
0.4
100
69
(2)
5
0
333
(3)
5
100
322
30
Figure 43: Time history of orbital maneuvers.
SPACECRAFT CHARGING
Spacecraft charging occurs as a result of the fact that at a given temperature, the electrons have a
much higher velocity than do the ions, and a potential bias builds up on spacecraft surfaces. This is a
well studied problem at geostationary orbit altitudes, and lesser so at LEO. A relevant study is described
from charging events and anomalies on DMSP F13 spacecraft by Anderson and Coons [Anderson, 1996],
for a satellite where unusual anomalies were experienced at 800 km. For the nominal ionospheric
densities, surface charging to a few volts is not unusual. The amount of charge collected is related to the
amount of conducted material exposed to the environment. For the DMSP F13 spacecraft, the high
latitude exposure to the auroral ionosphere caused an anomalous condition of very large potentials (100s
of Volts), which lead to discharge(s). In the case of DMSP, electron fluxes were nominally <105 cm-2s-1,
but exceeded 108 cm cm-2s-1 for periods of aurorally enhanced ionospheric exposure.
Spacecraft charging will be an issue to deal with. The deployed 20 m2 area of an active conductor
will be a very effective charged material regardless of direction to ram. For nominal ionospheric densities,
charging potentials would be larger than those of DMSP, and for enhanced ionospheres, CubeSail would
experience similar charging on its conductive film as well as to whatever the film is attached. Discharging
of charged surfaces could lead to single event upsets as well as potential component failure.
SUMMARY AND CONCLUSIONS
UltraSail has been proposed as a controllable way to deploy and fly very large solar sails and the
concept appears scalable. A pair of CubeSat nanosatellites are being assembled to deploy and fly a 20
m2 sail as a low-cost technology demonstration in low Earth orbit. Modeling suggests that a gravity-
gradient deployment is the best option, and laboratory experiments are being used to support the design.
ACKNOWLEDGMENTS
This effort is performed under NASA Contract NNX09CB37C. Les Johnson of NASA Marshall
Spaceflight Center is the contract monitor. We also wish to acknowledge important technical discussions
on sail film technology with Greg Farmer and Mark Johnson at NeXolve, Inc., Huntsville, AL.
REFERENCES
Anderson, P. C. and Coons, H. C., Spacecraft Charging Anomaly on a Low-Altitude Satellite in
Aurora, J. Spacecraft and Rockets, Vol 33, pp 734-738, 1996.
31
Bonfiglio, E. P., Oh, D., and Yen, C-W, Analysis Of Chemical, REP, And SEP Missions To The Trojan
Asteroids, AAS 05-396, 2005
Botter, T. et al., Structural Dynamics Of A Spin-Stabilized, Mast-Free Solar Sail Design, Advances in
the Astronautical Sciences. Vol. 124, No. II, pp. 1935-1954, 2006.
Botter, T, Coverstone, V. L. and Burton, R., Structural Dynamics of Spin-Stabilized Solar Sails with
Applications to UltraSail, Journal of Guidance, Control, and Dynamics, Vol. 31, No. 2, 402-413, 2008.
Brogan, W. L., Modern Control Theory, Englewood Cliffs: Prentice Hall, 1991.
Bryson and Ho, Applied Optimal Control, Hemisphere Publishing Corporation, 1975.
Burton, R. L., Coverstone, V. L., Hargens-Rysanek, J., Ertmer, K. M., Botter, T., Benavides, G., Woo, B.,
Carroll, D. L., Gierow, P., Farmer, G., and Cardin, J., UltraSail, Ultra-Lightweight Solar Sail Concept,
AIAA No. 2005-4117, 2005.
Cal Poly San Luis Obispo, CubeSat Design Specifications Revision 12,
http://cubesat.org/images/developers/cds_rev12.pdf, 2009.
Cal Poly San Luis Obispo, CubeSat P-POD Mark II Rendering,
http://cubesat.org/index.php/media/pictures/55-p-pod-mk-ii-renderings, 2010.
Chobotov, V., Orbital Mechanics, 3rd Edition, AIAA Education Series, 2002
Edelbaum, T. N., Propulsion Requirements for Controllable Satellites, ARS Journal, Vol. 31, 1079-
1089, 1961.
Edelbaum, T. N., Optimum Low-Thrust Transfer between Circular and Elliptic Orbits," Proc. Fourth
Nat. Cong. Appl. Mech., (ASME, N. Y.) pp. 137–141, 1962.
Ertmer, K., Design and Operation of a Thin-Film Vacuum Deployment Experiment for UltraSail
Concept Validation, MS Thesis, Dept. of Aerospace Engineering, University of Illinois, 2006.
Gaposchkin, E.M., Calculation of Satellite Drag Coefficients, Technical Report 998, Lincoln
Laboratory, MIT, Lexington, Massachusetts, 18 July, 1994.
Goldberg, D. A., Genetic Algorithms in Search, Optimization, and Machine Learning, Boston:
Addison Wesley Longman, Inc., 1989.
Hargens-Rysanek, J., The Dynamics and Control of the UltraSail System, Ph.D. Thesis, Dept. of
Aerospace Engineering, University of Illinois, 2006.
Hargens-Rysanek, J., Coverstone, V. and Burton, R., "Orbital Precession Via Cyclic Pitch For The
UltraSail System," Paper No. AAS 07-166, AAS/AIAA Space Flight Mechanics Meeting, Sedona, AZ,
January, 2007.
Hughes, P.C., Spacecraft Attitude Dynamics, John Wiley and Sons, 1986.
MacNeal, R. H., "The Heliogyro, An Interplanetary Flying Machine,” NASA Contractor's Report CR
84460, June 1967.
MacNeal, R. H., "Structural Dynamics of the Heliogyro", NASA-CR-1745A, 1971.
McInnes, C. R., Solar Sailing: Technology, Dynamics, and Mission Applications, Springer-Praxis
Publishing, Chichester, UK, 1999.
NASA Space Vehicle Design Criteria Guidance and Control, Spacecraft Aerodynamic Torques, Report:
NASA-SP-8058, Jan 1971.
Naval Research Lab NRLMSISE-00, http://www.nrl.navy.mil/content.php?P=03REVIEW105, 2003.
Psiaki, M., Magnetic Torquer Attitude Control via Asymptotic Periodic Linear Quadratic Regulation,
Journal of Guidance, Control, and Dynamics, vol. 24, pp. 386-394, March-April 2001.
32
Pukniel, A., Attitude Determination and Three-Axis Control System for Nanosatellites with
Magnetic Torque Actuation, M.S. Thesis, Dept. of Aerospace Engineering at the University of Illinois at
Urbana-Champaign, 2006.
Pukniel, A., et al., A Preliminary Study Of The Dynamics And Control Of The CubeSail Spacecraft,
AAS Paper No. 09-417, 2009 Astrodynamics Specialist Conference, Vol. 135, Advances in the
Astronautical Sciences, 2009.
Rowe, W.M., Sail Film Materials and Supporting Structures for a Solar Sail—a Preliminary Design,
Vol. IV, Jet Propulsion Laboratory, Pasadena, California, October, 1978.
Sinclair Interplanetary, 10mNm-sec Reaction Wheel Datasheet,
http://www.sinclairinterplanetary.com/10mNm-secwheel2009a.pdf, 2009.
Vulpetti, G., Johnson, L. and Matloff, G. L., Solar Sails A Novel Approach to Interplanetary Travel,
New York, Copernicus Books, 2008.
Wie, B., Solar Sail Attitude Control and Dynamics, Part 1, Journal of Guidance, Control, and
Dynamics, Vol. 27, No.4, July-August 2004.
Wright, J.L., Space Sailing, Gordon and Breach, New York, 1992.
