Content uploaded by Maria Grulich
Author content
All content in this area was uploaded by Maria Grulich on Dec 09, 2015
Content may be subject to copyright.
SMARD-REXUS-18: DEVELOPMENT AND VERIFICATION OF AN SMA BASED
CUBESAT SOLAR PANEL DEPLOYMENT MECHANISM
Maria Grulich1, Artur Koop1, Philipp Ludewig1, Johannes Gutsmiedl1, Johannes Kugele1, Thomas Ruck1, Ingo
Mayer1, Alexander Schmid1, and Karl Dietmann2
1Technische Universität München, Boltzmannstr. 15, 85748 Garching, Germany, list@smard-rexus.de
2Hochschule München, Lothstraße 34, 80335 München, Germany, list@smard-rexus.de
ABSTRACT
SMARD (Shape Memory Alloy Reusable Deployment
Mechanism) is an experiment for a sounding rocket de-
veloped by students at Technische Universität München
(TUM). It was launched in March 2015 on REXUS 18
(Rocket Experiments for University Students).
The goal of SMARD was to develop a solar panel hold-
down and release mechanism (HDRM) for a CubeSat us-
ing shape memory alloys (SMA) for repeatable actuation
and the ability to be quickly resettable.
This paper describes the technical approach as well as the
technological development and design of the experiment
platform, which is capable of proving the functionality of
the deployment mechanism. Furthermore, the realization
of the experiment as well as the results of the flight cam-
paign are presented.
Finally, the future applications of the developed HDRM
and its possible further developments are discussed.
Key words: SMARD; REXUS; TUM; CubeSat;
MOVE-II; SMA; HDRM.
1. INTRODUCTION
Since 2011 the Institute of Astronautics (LRT) and
the scientific student group WARR (Scientific Work-
group for Rocketry and Space Flight) have been de-
veloping a CubeSat called Munich Orbital Verification
Experiment-II (MOVE-II). Because the payload requires
at least 20 W of power, the satellite will be equipped with
four 210x80 mm solar panels. These panels will be in a
stored configuration during launch and will be deployed
in orbit, which makes an HDRM necessary.
Usually a single-shot system is used for such a mech-
anism. MOVE-II’s predecessor CubeSat First-MOVE
used a mechanism based on melting wires, which was
a single-shot system. This HDRM was launched and
tested on REXUS 4 by the student team VERTICAL from
TUM.[1] The testing and qualification of First-MOVE
showed that the ability to easily reset the HDRM should
be an important design consideration. The reusability
simplifies testing and reduces the time between tests and
the possibility of a potential failure during the assembly
procedure.
For this reason it was decided not to use the flight-proven
First-MOVE mechanism, but to instead develop a new
system using a NiTi-based SMA spring.
To ensure a successful operation of the mechanism dur-
ing MOVE-II’s mission, thorough testing of the mecha-
nism in all relevant environments is necessary. With high
vibrational stresses and g-loads during launch, as well
as the vacuum and milli-gravity environment during the
coast phase, a sounding rocket is ideal for this purpose.
The milli-gravity phase should be used to measure the os-
cillations of the solar panel. Therefore an accelerometer
and a magnetic rotary decoder are installed. A thermistor
measures the temperature near the SMA spring.
A camera (GoPro HERO3+ Black) documents the pro-
gression of the experiment, especially the deployment
and movements of the solar panel in milli-gravity.
For controlling and interfacing with the rocket in milli-
gravity, ERIS (Electronic Reliable Integrated System)
was developed. ERIS is based on the Compatible and Ex-
tendable REXUS Experiment Support Bus (CERESS), an
experiment bus for REXUS developed by four students
from TUM and flown on REXUS 13.[2]
It controls the experiment and provides data acquisition
and storage. The used software is based on National In-
struments LabVIEW. The ground station provides dis-
plays for all important signals and sensors. All sensor
data are transmitted to the ground via the REXUS teleme-
try interface.
2. EXPERIMENT OVERVIEW & DESIGN
The SMARD experiment is installed inside a standard
300 mm REXUS module. It consists of three structural
components mounted on a bulkhead as can be seen in
Fig. 1: the camera housing, the ERIS housing and the
frame carrying the experiment assembly.
The camera housing is made of aluminum with a glass
fiber reinforced polymer (GFRP) back plate. The on-
board computer, ERIS, consisting of three printed circuit
boards (PCBs) is mounted inside an aluminum housing
Figure 1. Overview SMARD Experiment
which provides stability and shielding.
The experiment itself consists of the mechanism un-
der test and a generic two-unit CubeSat solar panel
dummy loosely based on preliminary requirements from
the MOVE-II mission.
Figure 2. Solar Panel
The experiment frame is a simple aluminum design pro-
viding the necessary stiffness for the experiment assem-
bly to survive the launch on a REXUS rocket as well as a
mounting point for the panel magnetic rotary sensor.
The panel is made of carbon fiber reinforced polymer
(CFRP) and carries thin aluminum sheets representing
the solar cells. The panel is connected to the mounting
frame through two hinges, which contain springs for the
deployment of the panel.
The panel’s resting position is defined by stoppers on the
hinges at an angle of 135◦from the closed position.
A hook mounted on the solar panel is the connection to
the HDRM. The HDRM secures the hook and therefore
keeps the solar panel in the closed position during launch.
In this configuration the HDRM as well as the hinges
have to withstand all launch loads up to 17 g.
After launch, the HDRM is activated and the panel is
opened by the spring hinges.
Figure 3. Hinge
2.1. HDRM
The heart of the HDRM is the SMA spring. SMAs are
able to "remember" their original shape. This effect is due
to a solid-state phase transformation between two crys-
talline structures in the metal. In the cold state, the crys-
talline structure is detwinned martensite. In this phase
the SMA can be deformed as needed. When the alloy is
heated up the crystalline structure transforms to austen-
ite, which means it returns to its pre-deformed shape.[3]
SMAs can have a reversible stretching of 4-6% and can
be used for at least 1,000 cycles. They exist in a lot of
different forms, for example: tubings, wires, sheets, foils
and springs.[3]
The temperature at which the transformation from
martensite to austenite starts is called activation or trans-
formation temperature and can be adjusted by slight
changes in alloy composition and through heat treatment.
In the NiTi alloys, for instance, it can be changed from
above 100◦C to below -100◦C.[4]
The activation temperature of the SMA spring used for
SMARD is about 85◦C. When the SMA cools down it
transforms into a twinned martensite phase, until it is de-
formed again mechanically, which resets the metal of the
used spring to its starting configuration, as illustrated in
Fig. 4.
Figure 4. Solid-State Phase Transformation of Shape
Memory Alloys [3]
The SMA spring used for the HDRM has a force of 6 N
and a stroke of 6 mm. The outer diameter is 4.47 mm and
the used wire diameter is 0.735 mm.
The HDRM consists of five parts: a housing, an alu-
minum slider, an insulation plate, a mechanical spring,
as well as an SMA spring.
The slider is placed in the housing, while both springs
are attached to the slider on one end and to the housing
on the other end, as can be seen in Fig. 5. Both springs
operate as tension springs. The whole HDRM is then
mounted onto the experiment frame with an insulation
plate made of glass fiber between the HDRM and the alu-
minum (Fig. 6 insulation plate not shown).
The HDRM’s main task is to either lock the solar panel
hook in position or release it.
Figure 5. HDRM
Figure 6. Positioning of the HDRM
This is done by the slider, which can move between two
positions: closed (during launch) and open.
Closed is the default position, when the HDRM is not ac-
tivated (Fig. 7).
In order to contract the SMA spring, the activation tem-
perature of about 85◦C has to be reached. By applying an
electrical current of 1.9 A at a voltage of 3.3V for 10s the
SMA spring heats up to the activation temperature due to
the ohmic resistance of the spring wire itself.
With a contraction force of about 6 N the SMA spring
overcomes the opposing mechanical spring force and
therefore moves the slider into the open position (Fig. 8).
Figure 7. HDRM closed
Figure 8. HDRM open
2.2. Electronics Segment Overview
The electronics segment mainly consists of a measure-
ment and control system called ERIS, sensors to measure
experiment data and the activation of the HDRM.
ERIS is based on CERESS.[2] ERIS uses the newest
sbRIO board (sbRIO-9626), by National Instruments for
data storage and experiment control, using its integrated
SD-card and RS-485 interface.
It has smaller dimensions and lower mass than the one
used in CERESS.[5] The experiment is controlled by
ERIS which acquires data from all sensors and stores the
data on an on-board SD-card. Additionally, important
parts of the data are sent down to the ground station using
the CERESS telemetry protocol and the REXUS SM data
link.
2.2.1. Data Acquisition System
The sbRIO-9626 provides a powerful field-
programmable gate array (FPGA) with two million
gates as well as a 400 MHz real-time central processing
unit (CPU). Furthermore 110 digital I/Os and 32 ana-
logue channels are integrated.[5] The acquired data are
partially sent down to the ground station with the RS-485
interface while the rocket is in flight.
The sbRIO is connected to two Power and Sensor Inter-
face Boards (PSIBs). These boards provide connections
to all sensors and the REXUS SM. The REXUS Service
Module (SM) supplies all experiments with power. [6]
Since its 28 V are unregulated and too high for the
components used in the experiment a voltage converter
is needed that is also located on the PSIBs. They provide
regulated power for all systems including the sbRIO.
Furthermore they include a voltage converter for the
HDRM. They provide regulated power for all systems
including the sbRIO. Furthermore they include a voltage
converter for the HDRM.
This voltage converter has an output voltage of 3.3 V and
can deliver up to 5 A of current which is approximately
the range of power, which a CubeSat can provide for the
HDRM.
2.2.2. Sensors
Several sensors are used to measure the deployment of
the solar panel and to monitor the temperatures of the
electrical components.
The magnetic rotary encoder chip tracks the opening an-
gle of the solar panel at a sample rate of 10 kHz with
a resolution of 12 bit over a full revolution of 360◦.
Two accelerometers (type LIS331H) with a range of
±24 g and the same 10 kHz sample rate are used to track
panel oscillations while opening. Additionally two NTC-
thermistors are used to measure the heating of the SMA
spring and the surrounding components.
2.3. Software Design
The software for SMARD is part of the ERIS system and
was programmed, based on the CERESS software, using
LabVIEW 2013. The main functions of ERIS can be de-
scribed as follows:
Timeline Execution of Events: Although the REXUS
SM provides the experiments with three user defin-
able signals, ERIS utilizes an internal timeline in or-
der to trigger more events (e.g. restarting the cam-
era). The three signals are: Start Of Experiment
(SOE), Start Of Data Storage (SODS) and Lift-Off
(LO).[6]
The timeline is executed on the real-time (RT) CPU
of the sbRIO. When a certain event should occur, the
timestamper of the FPGA sends an interrupt.[2]
Timestamping: Using the FPGA of the sbRIO-9626,
ERIS utilizes two timers: one which is started dur-
ing the booting sequence of the system and the other
that is started after LO in order to be able to com-
mence the timeline at exactly defined times. Both
timestampers have an accuracy of up to 1 ms.
Telemetry Generation: Telemetry packages, which are
generated on the FPGA and sent by the RT CPU, are
used to send data to the ground station as a backup.
The message format used in ERIS, like CERESS,
includes a 40 bit header, the relevant data, and for
verification of the sent data a checksum and cyclic
redundancy checksum.
Event Logging: After an event a log file is generated
and stored on board on the internal non-volatile stor-
age of the sbRIO-9626. It contains all state changes,
signal changes and file information, which prevents
unwanted data loss due to file overwriting.
Signal Detection and Handling: The signals provided
by the REXUS SM (SOE, SODS and LO) are de-
tected by ERIS and observed in case of a change of
the signals. Since the LO signal is hardware driven,
a debouncing of the signal is necessary. For this pur-
pose, if the signal is true, the software checks the
signal for a time span of 200 ms. Every 10 ms a sam-
ple is taken and only when all samples are true, the
signal will be accepted as received by the OBDH.
Data Acquisition: There are two types of data acquisi-
tion implemented in ERIS: the digital data acquisi-
tion and the analogue data acquisition. The temper-
ature and maintenance data, such as voltage and cur-
rent of the supply lines, are analogue signals, which
do not need a certain protocol. The accelerometers
are connected via serial peripheral interface (SPI)
bus. Voltage and current of the SMA spring power
supply and the voltage across the SMA spring while
it is being activated is measured using an analogue-
to-digital converter (ADC) chip connected via SPI
to the sbRIO. The current running through the SMA
spring while it is being activated is measured using a
Hall effect based current sensor, which is connected
to an analogue channel of the sbRIO.
Data Storage: The data generated by the data acquisi-
tion system are stored on the on-board 2 GB SD-card
with a possible maximum rate of 2 MBps. ERIS uses
the File Allocation Table (FAT) file system provided
by the SD-Card Module.
The binary data are stored in one file on the SD-card,
since different acquisition rates are used (10 kHz for
the magnetic rotary encoder, 1 kHz for the acceler-
ation sensors and 1 Hz for the remaining analogue
sensors).
Camera and Mechanism Control: A digital signal on a
pin of the GoPro connector activates the camera and
a signal on another pin initializes the recording and
storage of the images. The camera is deactivated
after the experiment has ended in order to ensure a
safe storage of the video file.
The mechanism on the other hand is also controlled
by a digital signal to the interface boards. These then
apply power to the mechanism, thereby activating it.
2.4. Ground Segment
The ground station (GS) is implemented in LabVIEW
2013 and SQL.
Since the main purpose of the ground segment is to re-
ceive real-time data from ERIS, it has only been slightly
modified from the original CERESS GS software.[2]
The SMARD ground segment consists of the ground sta-
tion, including a server and clients, which interface with
the ESRANGE ground segment. The SMARD ground
segment is connected to ERIS via the RS-232 TM/TC
connection provided by the ESRANGE ground segment
system. The main functions of the ground station are:
•Provide status information of the experiment (accel-
eration, temperature, voltage, errors, etc.)
•Verify experiment success
•Data backup in the unlikely event of recovery failure
The heart of the ground station is a MySQL database that
handles the distribution of telemetry (TM) packages and
provides access for multiple clients, such as the scientific
display, which shows the acceleration values of the sen-
sors, by using the Open Database Connectivity-Protocol
(ODBC) over LAN. Therefore further clients can easily
be added without the need of a time-consuming system
reconfiguration.
The database is fed by a LabVIEW program (TM re-
ceiver) running on a Windows PC, where the RS-232
connection is located. This TM receiver handles the
low-level RS-232 serial communication functions such as
byte-wise reading from the serial port buffer.
In order to handle the incoming serial data, a TM decoder
program checks the packages for errors and reads the re-
port ID of the package to identify the data type. Based
on this, the appropriate decoder program splits the data
stream according to the decoding definition.
These separated data packages (binary format) are then
forwarded to an SQL-handler which sends them to the
MySQL database via LAN. Multiple clients can now con-
nect to the database and asks for data, which is returned
to the client via SQL. These data are then parsed by the
client and the results are shown on the displays.
3. FLIGHT RESULTS
The EuroLaunch RX-18 housekeeping data show that
REXUS 18 lifted off at 13:29 UTC on the 18th of March
2015 from ESRANGE, reaching its apogee of 81.7 km at
140 s into the flight.
During the launch SMARD was exposed to a maximum
of 16.8 g. A yo-yo de-spin to initiate the milli-gravity
phase was unsuccessful, resulting in a continued roll rate
of 2.7 Hz and thus, due to the centrifugal acceleration of
this rotation, preventing the rocket from entering a state
of weightlessness. Upon reentry, a flat-spin phase around
all three axes of the rocket, exerted strong and quickly
alternating forces on the experiment for roughly 150 s.
3.1. HDRM and Panel Movements
Fig. 9 shows the rotation of the panel around its axis. In-
stead of swinging open, the solar panel remained closed
due to the centrifugal forces, which were caused by the
continued spinning of the rocket. However, a small ori-
entation change of the panel occurred as the HDRM was
activated shortly after the 600 s mark. This is due to
the fact that the HDRM performed nominally and the
panel was released from the mechanism. This thesis is
also supported by accelerometer data in Fig. 10, where a
peak upon HDRM activation shows that the panel slightly
moved. Finally, in the GoPro footage the release can be
clearly seen.
Strong movements (Fig. 9) and accelerations (Fig. 10)
can be observed from 720 s onward. Here the panel
is opened and closed rapidly by alternating centrifugal
forces due to the rocket being in flat spin during reentry.
This behavior prevents evaluation of the opening move-
ment of the solar panel as it would have been in milli-
gravity.
Since the panel did swing open and closed, when the di-
rection of forces changed, and the hook as well as the
HDRM itself were undamaged after the flight, it can
be said that the SMARD mechanism itself performed as
planned and released the panel upon activation.
3.1.1. SMA Spring
As soon as the liftoff occurred the SMA spring started
to heat up as can be seen in Fig. 11 (page 7), resulting
from the heating of the rocket itself. However, one can
still clearly see a change in temperature gradient at the
time the spring actuation occurred. Concerning the maxi-
mum temperature of the SMA spring, a precise statement
Figure 9. Magnetic Rotary Decoder Data of the Solar Panel Movements
Figure 10. Acceleration of Panel Z-Axis in Bulkhead Coordinates (Differential)
cannot be made due to the lack of direct contact between
the sensor and the SMA spring. The maximum tempera-
ture of a component was the bulkhead with about 65◦C,
which is less than the activation temperature of 85◦C for
the SMA spring.
Figure 11. Temperature Curves of sbRIO (blue), Bulk-
head (red), Traco (yellow) and the SMA Spring (purple)
In Fig. 12 you can see the voltage across the HDRM.
The used SMA spring driver section is set to a voltage of
3.3 V. The spring driver section should regulate the volt-
age to this fixed level, but it is normal that due to errors in
the feedback loop the voltage can vary a little bit. Also,
the voltage stays below to the nominal 3.3 V due to the
voltage drop in the cables of the HDRM. The used SMA
spring driver can deliver up to 5 A of current.
The maximum current of 1.95 A (Fig. 13) and the max-
imum power draw of 6.15 W of the SMA were slightly
less than expected but within the constraints of the
MOVE-II electrical power system, proving the usabil-
ity of an SMA-based mechanism on CubeSats. Due to
the rising temperature and the associated increase of the
springs’ ohmic resistance, it can be observed that both the
power and current decrease while the mechanism is being
activated (Fig. 13).
The MOVE-II CubeSat design currently employs an EPS
board from Clyde Space that is capable of delivering
a maximum power of 30 W depending on battery load-
ing state. This shows that the developed HDRM could
be used without any problems regarding the power con-
sumption.
The decrease in current and power during activation tells
us, that the resistance of the HDRM increases. On the
one hand this effect results from the rise in temperature
of the SMA material and on the other hand the resistance
also increases because of an increase of contact resistance
between the conductive parts inside the HDRM during
movement.
Figure 12. SMA Voltage
Figure 13. SMA Current
3.1.2. Electronics
Throughout the flight, SMARD’s electronic boards and
sensors all performed nominally and collected data which
was stored and, in parts, sent down to the ground station.
Fig. 11 shows the temperatures of different components
over time. By comparing the curves of the bulkhead tem-
perature and the main DC/DC converter temperature (la-
beled "Traco" in Fig. 11), one can clearly see how well
insulated the electronics boards are from the bulkhead:
especially for the Traco, the temperature rise is primarily
due to its self-heating since the rate of heating does not
increase at liftoff.
3.2. Post Flight Tests
After the experiment was recovered and returned to the
SMARD team it turned out the panel itself took no dam-
age from the strong "flapping" movements during reen-
try. In contrast the hinges took severe damage, including
plastic deformation of the lever, making the panel hard to
move around its rotation axis. Since the damage results
from the harmful conditions of reentry, it does not impair
the hinge’s future use for MOVE-II.
The HDRM itself including the SMA spring was tested
and activated after the launch and still worked nominally,
proving the SMARD-mechanism’s durability and suit-
ability for a CubeSat and the multitude of tests preceding
its launch.
4. CONCLUSION
In summary the team SMARD developed a fully func-
tional HDRM for CubeSats based on SMA. It was suc-
cessfully verified on board a sounding rocket. An impor-
tant goal of the experiment was to measure and observe
the oscillations of the solar panel during milli-gravity.
Since the REXUS 18 mission unfortunately did not have
a milli-gravity phase, these measurements could not be
made. The estimated success can be split into three cat-
egories based on the experiment objectives. Success val-
ues have been assigned to each category with a total of
100 % meaning complete experiment success:
•Functional verification of the mechanism on its first
flight: 60 %.
•Visualization of solar panel deployment: 10 %.
•Measurement of solar panel movement in milli-
gravity: 30 %.
The mechanism functioned as expected and the camera
visualized the panel’s oscillations. The measurement of
the solar panel’s movements using the magnetic rotary
decoder and the accelerometer worked as expected. Due
to the lack of milli-gravity no data related to the be-
haviour of the solar panel in a reduced gravity environ-
ment could be generated. All in all the resulting success
of the flight for SMARD is 70 %.
5. OUTLOOK
The heart of the SMARD project was the development of
a HDRM for CubeSats. This mechanism will be further
developed within the MOVE-II project.
Since the SMARD mechanism proved to work success-
fully during the launch and flight, with even harder con-
ditions than expected for an orbital mission, it has made
a big step towards being used on the MOVE-II satellite.
The evaluated flight data are already being used at TUM’s
Institute of Astronautics to further scale down the HDRM
within the scope of a bachelor thesis.
The aim of the bachelor thesis is to improve the electrical
conductivity as well as the size of the mechanism and to
consider a protection against possible resonance effects.
ACKNOWLEDGEMENTS
The authors would like to thank all the institutions and
companies that supported the SMARD project, specifi-
cally:
•DLR/DLR-MORABA
•ZARM
•SSC
•SNSB
•RYMDTYRELSEN
•ESA
•National Instruments
•Ingpuls
•Gutekunst Federn
•DAAD
•Bund der Freunde der TUM
•WARR
•TUM/LRT
REFERENCES
[1] Olthoff C., Purschke T. R., Winklmeier, R., Czech,
M.,2010, Testing of Critical Pico-Satellite Systems on the
Sounding Rocket Rexus-4. In: Rainer Sandau, Hans-Peter
Roeser und Arnoldo Valenzuela (Hg.): Small Satellite Mis-
sions for Earth Observation. Berlin, Heidelberg: Springer
Berlin Heidelberg, S. 257-266.
[2] Schmitt A., Bugger D., Friedl C., Althapp S.,2014, Provid-
ing a standard support bus for sounding rocket experiments
with the CERESS experiment.
[3] Schiedeck F.,2009, Entwicklung eines Modells für For-
mgedächtnisaktoren im geregelten dynamischen Betrieb.
[4] Nitinol for Medical Devices. Use the superelasticity and
shape memory properties of Nitinol in your next medical ap-
plication. http://www.jmmedical.com/nitinol.html.
[5] National Instruments Corp., NI Single-Board RIO Embed-
ded Control and Acquisition Devices-NI-sbRIO-96XX.
[6] Fittock M., and Inga M.,2012, REXUS User Manual. Euro-
Launch.
