See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/363862530

Low-Cost Attitude Determination and Control System of the Student-Built

3U+ CubeSat SOURCE

Conference Paper · September 2022

CITATIONS READS

0 276

11 authors, including:

Nadim Maraqten Paul Jannik Haufe

Universität Stuttgart Universität Stuttgart
4 PUBLICATIONS   1 CITATION    3 PUBLICATIONS   2 CITATIONS   

SEE PROFILE SEE PROFILE

Martin Zietz Alexander Wagner

Universität Stuttgart Universität Stuttgart
2 PUBLICATIONS   1 CITATION    1 PUBLICATION   0 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Stratospheric Balloon View project

Space Mobile Robots View project

All content following this page was uploaded by Nadim Maraqten on 27 September 2022.

The user has requested enhancement of the downloaded file.

 73rd International Astronautical Congress (IAC), Paris, France 18-22 September 2022.
Copyright 2022 by Mr. Nadim Maraqten. Published by the IAF, with permission and released to the IAF to publish in all forms.

IAC-22,E2,3-GTS.4,10,x70167

Low-Cost Attitude Determination and Control System of the Student-Built 3U+ CubeSat SOURCE

Nadim Maraqtena∗ , Paul J. Haufea , Martin Zietza , Alexander Wagnera , Christopher Vogta , Luc Lauera ,
Nicolas A. Probsta , Justus Golla , Steffen Gaisserb , Robin Schweigertb , Sabine Klinknerb

a
Small Satellite Student Society University of Stuttgart (KSat e.V.), Stuttgart, Pfaffenwaldring 29, 70569 Stuttgart
maraqten[at]ksat-stuttgart.de
b
Institute of Space Systems, University of Stuttgart, Pfaffenwaldring 29, 70569 Stuttgart
∗
Corresponding Author

Abstract

Due to a substantial decrease in cost within the past decades, the realm of space has become accessible to a myriad
of satellite missions. To contribute to this trend and make space more affordable, this paper presents the low-cost
Attitude Determination and Control System (ADCS) of the student-built 3U+ Stuttgart Operated University Research
CubeSat for Evaluation and Education (SOURCE). The system can be used as a blueprint for low-cost ADCS designs
and developments of future missions. The ADCS on-board SOURCE is vital for the success of its scientific missions:
meteor and Earth observation, in-situ demise analysis and space-verification of Commercial Off-The-Shelf (COTS)
components. SOURCE is a collaborative project between the Small Satellite Student Society of the University of
Stuttgart (KSat e.V.) and the University of Stuttgart-Institute of Space Systems (IRS). It is supported by the "Fly Your
Satellite! 3" programme of the European Space Agency (ESA).
SOURCE relies on its ADCS primarily for sun-pointing of the solar panels and accurate attitude determination for
the scientific payloads. The CubeSat conducts measurements of atomic oxygen, pressure and heat flux. These are
especially important for atmospheric measurements during the final mission phase: re-entry. The satellite’s attitude
determination is provided by data from photodiodes used as sun sensors, Magnetometers (MGMs), Global Navigation
Satellite System (GNSS) receivers and gyroscopes, most of which are inexpensive COTS components originally de-
veloped for Earth applications. Their data is used in self-developed attitude and orbit determination laws, that yield
the current state and the desired actuation. The CubeSat’s actuation is entirely based on Magnetorquers (MGTs),
which provide two-axis control. The MGTs are self-built at exceptionally low cost. This in-house development allows
for adaption to the satellite instead of vice versa — as would be the case with commercial ones. For verification
and validation all systems undergo in-depth testing including functional, thermal-vacuum, shaker and electromagnetic
compatibility tests.
The self-developed ADCS algorithms are examined using ESA’s Generic AOCS GNC Techniques and Design Frame-
work for FDIR (GAFE). Additionally, the satellite’s dynamic behaviour and the performance of its ADCS during the
early re-entry phase are analysed using a self-developed aerodynamic model that is based on drag coefficients from
numerical free molecular flow simulations. The conducted tests of the developed ADCS validate the system’s integrity,
taking the SOURCE CubeSat one step closer to its launch and the realization of its scientific mission. The developed
system serves as an inspiration for future low-cost ADCS.
Keywords: ADCS, CubeSat, SOURCE, Low-cost, COTS, Re-entry.

Nomenclature Kω Gain matrix for rate damping torque

Aref (m2 ) Aerodynamic reference area
Lref (m) Aerod. reference length
α (°) Angle of attack
β (°) Sideslip angle
ωref (rad s−1 ) Reference angular rate vector
ω (rad s−1 or ° s−1 ) Angular rate vector
δ (rad) Angle between sun vector and reference direction
B (T) Magnetic field vector
CF (-) Aerod. force coefficient vector
CM (-) Aerod. torque coefficient vector
FAdrag (N) Aerod. force vector in aerod. coordinates
Mω (N m) Rate damping torque for higher modes
Mal (N m) Alignment torque for higher modes
MBdrag (N m) Aerod. torque vector in body coordinates
M (N m) Commanded torque for higher modes
XB
A (-) Aerod. X-axis in body coordinates
XB
B (-) The satellite’s X-axis in body coordinates
B
YA (-) Aerod. Y-axis in body coordinates
B
ZA (-) Aerod. Z-axis in body coordinates
∆∥B (rad s−1 ) Angular rate error parallel to the magnetic
field vector
∆⊥B (rad s−1 ) Angular rate error perpendicular to the
magnetic field vector

IAC-22,E2,3-GTS.4,10,x70167 Page 1 of 14
 73rd International Astronautical Congress (IAC), Paris, France 18-22 September 2022.
Copyright 2022 by Mr. Nadim Maraqten. Published by the IAF, with permission and released to the IAF to publish in all forms.

∆ (rad s−1 ) Total angular rate error in relation to the mag- by the ADCS. Navigation algorithms generate attitude in-
netic field vector formation based on the sensor readings. Guidance algo-
m (N m) Commanded torque Detumble Mode rithms detect the attitude error which is fed into the atti-
sref (-) Satellite reference direction in geometry frame tude control laws. The target attitude is achieved by com-
s (-) Sun direction in geometry frame manding the available actuators.
B
vrel (m s−1 ) Relative airstream velocity in spacecraft co- Spacecraft without ADCS face the same problem as the
ordinates first human-made satellite Sputnik 1 in 1957: They ro-
ω⊥B (rad s−1 ) Angular rate perpendicular to the magnetic tate arbitrarily, making many operations a real difficulty:
field vector conducting directed scientific measurements of Earth, tak-
ρ (kg m−3 ) Air density ing photos, pointing the antenna at a ground station, or
kω Gain factors for rate damping torque pointing the solar panels at the sun. These reasons ignited
kal Gain for alignment torque the increasing development and establishment of ADCS
in the late 1970s [1]. Nowadays, ADCS can be found in
Acronyms/Abbreviations most spacecraft [2].
AC Air-core
1.1 The SOURCE CubeSat Mission
ADCS Attitude Determination and Control System
C Cylindrical For the 3U+ CubeSat mission SOURCE the ADCS is
COTS Commercial Off-The-Shelf vital for a variety of scientific and operational reasons:
DSMC Direct Simulation Monte Carlo meteor and Earth observation, sun-pointing and in-situ
ECSS European Cooperation for Space Standardization demise analysis [3]. During reentry below 200 km, mea-
EM Engineering Model surements of temperature, heat flux, pressure and atmo-
ESA European Space Agency spheric oxygen will be collected and downlinked [4]. Fur-
FDIR Failure Detection, Isolation, and Recovery ther aims comprise technology demonstration and space-
FIPEX Flux Phi Probe Experiment qualification of COTS components [5, 6], which are di-
FLP Flying Laptop rectly linked to the low-cost aspect. Low-cost is also
FM Flight Model achieved by a magnetorquer-only actuation, instead of
GAFE Generic AOCS GNC Techniques and Design combining them with reaction wheels or thrusters. Fur-
Framework for FDIR thermore, one of the main objectives of the mission is ed-
GNSS Global Navigation Satellite System ucation [3, 7]. The mission is a student-led joint venture
HKB Housekeeping Board between the Small Satellite Student Society of the Uni-
IRS University of Stuttgart-Institute of Space Systems versity of Stuttgart (KSat e.V.) and the Institute of Space
KSat e.V. Small Satellite Student Society of the Univer- Systems of the University of Stuttgart (IRS). The mission
sity of Stuttgart moreover receives support from ESA within the "Fly your
LEO Low Earth Orbit Satellite! 3" programme and is currently in phase D, ex-
MEKF Multiplicative Extended Kalman Filter pecting to launch in 2023. A render of the CubeSat is
MeSHCam Meteor, Star and Horizon Tracking Camera shown in Fig. 1.
MGM Magnetometer
MGT Magnetorquer
NASA National Aeronautics and Space Administration
OBC Onboard Computer
PCB Printed Circuit Board
QM Qualification Model
QUEST Quaternion Estimator
SOURCE Stuttgart Operated University Research Cube-
Sat for Evaluation and Education
SuS Sun Sensors
TBD To be determined
TC telecommand Fig. 1. Render image of 3U+ CubeSat SOURCE with de-
TVac Thermal Vacuum Chamber ployed solar panels

1. Introduction
Increasingly complex space missions necessitate ac- 1.2 Objective of the Paper
curate knowledge about the spacecraft’s orientation and This paper presents the low-cost magnetically actu-
position, as well as their control. These tasks are realised ated ADCS of the SOURCE CubeSat to serve as an inspi-

IAC-22,E2,3-GTS.4,10,x70167 Page 2 of 14
 73rd International Astronautical Congress (IAC), Paris, France 18-22 September 2022.
Copyright 2022 by Mr. Nadim Maraqten. Published by the IAF, with permission and released to the IAF to publish in all forms.

ration and blueprint for low-cost ADCS of other CubeSat
missions. As a result of the paradigm change in space
accessibility through drastically decreased launch costs
[8], the actual system costs become increasingly impor-
tant. Instead of hoping for novel disruptive technologies,
e.g. ferrofluid based attitude control [9], certain strategies
related to already existing and well-established technolo-
gies can be employed. If successfully implemented, these
can be mission enablers, especially for small satellites and
missions with limited budgets.
The main strategies for cost-effectiveness exploited by
the SOURCE team are the use of COTS components
and building in-house wherever possible (e.g. magne-
torquers). It should be kept in mind that the pragmatic
use of COTS components must not necessarily imply an
impairment of functionality. National Aeronautics and
Space Administration (NASA)’s Mars helicopter Ingenu-
ity, for example, utilised a cellphone grade accelerometer
and gyroscope [10]. Similar components were also found
to be suitable for SOURCE. All components included in
the SOURCE ADCS are subject to extensive functional,
thermal-vacuum, shaker and electromagnetic interference
tests, qualifying them for use. If not already available,
space heritage will be conferred during the mission. This
makes the components attractive for use in future satel-
lites.
the prograde direction for 120 s for each of their measure-
ments. During the reentry science phase, the design does
not impose any requirements on the controllability of the
satellite. However, various payloads require reliable atti-
tude and orbit determination to give context to their data
throughout the mission, including reentry. The system
modes will be further elaborated in Section 3.
Because of size, mass and budget constraints, no reaction
wheels or thrusters are available and an architecture with
only magnetorquers was chosen. These are constrained in
size and power consumption by the overall design. This
also means, that the pointing requirements are only valid
as long as the atmospheric disturbance torques do not
overpower the MGTs. This is expected to occur below
∼240 km [11].
These facts translate to the key requirements:
• Compatibility with separation tip-off rates up to
90 ° s−1 per axis
• Detumbling within 5 hours of deployment
• Reliable coarse sun pointing in Safe Mode (20°)
• Attitude determination accuracy better than 5° in Idle
Mode and during reentry measurements

1.3 Outline • Altitude determination better than 1000 m until mis-

To define the framework conditions, the paper gives sion end
an overview on the SOURCE-specific Requirements and
System Modes. According to these, the ADCS Compo- • Pointing accuracy better than 20° in Idle Mode
nents are selected and the associated Hardware Testing
is presented. Ensuing, the Attitude Determination and
• Angular rates ∥ω∥ < 1 ° s−1 during pointing
Control Laws are discussed. The concluding section is
the Analysis of the ADCS performance during Re-Entry,
• MGT dipole moment higher than 0.3 A m2 on each
which is especially important for one of the major scien-
axis, more desirable within size and power con-
tific objectives of the CubeSat: demise investigations [4].
straints

2. Requirements
SOURCE is a 3U+ CubeSat with a mass of 5 kg. The
satellite’s dimensions can be taken from Fig. 2. As a con-
sequence of its directional solar array, it is mission critical
to detumble it and point the solar panels to the sun before
the battery runs out. As the launch provider is still un-
known, high separation tip-off rates have to be assumed.
After reaching a stable state ("Safe Mode"), the ADCS
can switch to the operational "Idle Mode" and the require-
ments are mainly driven by the payload. The Meteor, Star
and Horizon Tracking Camera (MeSHCam) needs to be
pointed at Earth for observations with low enough angu-
lar rates to avoid motion blur, but does not need to track Fig. 2. Simplified model of 3U+ CubeSat SOURCE with
specific targets. Flux Phi Probe Experiment (FIPEX) sen- dimensions
sors, used to detect atomic oxygen, must be pointed in

IAC-22,E2,3-GTS.4,10,x70167 Page 3 of 14
 73rd International Astronautical Congress (IAC), Paris, France 18-22 September 2022.
Copyright 2022 by Mr. Nadim Maraqten. Published by the IAF, with permission and released to the IAF to publish in all forms.

3. System Modes
A−0 ACS−Off
SOURCE will use four main modes of operation: De-
TC TC
tumble, Safe, Idle and possibly a Pointing Mode. As the Scheduled
OR ω < ω_krit for 1500s
only attitude actuators on board are magnetorquers, the TC
OR Event
A−1 Detumble A−2 Safe
ω > ω_krit for 300s
main differences are what sensors are used in each mode OR
TC OBC OBC
to determine the satellite’s rotation rates and attitude. The Command Command

available sensors are Sun Sensors (SuS), Magnetometers

malfunction of
Gyro or GPS
TC
(MGMs), gyroscopes and GNSS receivers. An overview
malfunction of SuS
of components used in each mode can be found in Table 1. A−4 ACS−Idle
Transitions between the different modes can be triggered OBC
by telecommands (TCs) from the ground, specific events command
OBC
Command
and malfunctions or autonomously by the Onboard Com- OBC
Command
puter (OBC). The logic for this is shown in Fig. 3.
A−5 Re−Entry
After deployment from the launcher, SOURCE will enter

malfunction of
OBC

Gyro or GPS
Detumble Mode and reduce its rate of rotation to an ac- Command

ceptable value using the MGM data in a B-Dot controller.

malfunction of SuS
At this point it will enter the Safe Mode for commission- A−3 Pointing
(optional)
ing, pointing the solar panels at the sun using a minimum
of active components (SuS, MGM). The satellite may fall Fig. 3. Schematic of the control mode transitions in
back to Safe Mode in case of malfunctions. SOURCE’s ADCS
Nominal operations are performed in Idle Mode, which
also points the solar array towards the sun, but uses all
available attitude determination sensors. This allows atti- mentioned in Tab. 1. An overview of the incorporated
tude control even during eclipse. For Earth observation, ADCS components is given in Fig. 4.
SOURCE will stay in Idle Mode but change the target at-
titude from pointing the solar panels (+X) towards the sun
to pointing the MeSHCam (-X) to Earth. An additional
fine inertial pointing mode with an improved control law
is still being investigated to enhance payload measure-
ments.
Towards the end of the mission, when atmospheric distur-
bances exceed the MGTs capabilities, the system will be
switched to Reentry Mode. Currently, there are two com-
peting options for this mode under investigation, either
to only perform attitude determination, with the MGTs Fig. 4. Primary ADCS components (colored) flown on
switched off, or to utilize the detumble control law to 3U+ CubeSat SOURCE (transparent): sun sensors (blue),
achieve the dynamic stability described in Section 7. GNSS antennas (green), Housekeeping Boards including
magnetometers, gyroscopes, sensor readouts and MGT
control (red), magnetorquers (orange and yellow); depic-
Table 1. Components used in each ADCS mode tion without solar panels
Mode SuS MGM Gyro GNSS MGT
Detumble X X
Safe X X X 4.1 Sensors
Idle X X X X X To ensure pertinent measurements of the systems state
(Pointing) X X X X X at any point during the mission, the following sensors
Reentry X X X X TBD were chosen:

4. Components
The proper selection of the system components is key
since they determine the overall capabilities of the system
architecture and play a major role in the cost-effectiveness
of the system. To meet the requirements presented in Sec-
tion 2, SOURCE uses the sensors and magnetic actuators
4.1.1 Magnetometer
To detect the direction and intensity of the local mag-
netic field, three RM3100 magnetometers are used on
SOURCE. Furthermore, they are used during Detumble
mode, to determine the satellites angular velocity. To
avoid aliasing effects of rotation rates of up to 90 ◦ s−1 ,
a sampling frequency of 10 Hz was chosen.

IAC-22,E2,3-GTS.4,10,x70167 Page 4 of 14
 73rd International Astronautical Congress (IAC), Paris, France 18-22 September 2022.
Copyright 2022 by Mr. Nadim Maraqten. Published by the IAF, with permission and released to the IAF to publish in all forms.
4.1.2 Gyroscope
The measurement of angular rotation rate is performed
using two MEMS gyroscopes. Due to its small size and
the low price, the three axis Bosch BMI 270 was cho-
sen. This model provides a digital resolution of 0.004
◦ −1
√ −1
s LSB−1 and noise density of 0.008 ◦ s−1 Hz [12].
It thereby outperforms other comparable models.
In addition, this model was chosen by the EIVE mission,
also from the University of Stuttgart, allowing to share
test results and control algorithms.
4.1.3 Global Navigation Satellite System
As described in Section 2, reliable position informa-
tion is required until the end of the mission, both for pay-
load context and attitude determination. Because of this,
SOURCE uses a hot-redundant GNSS consisting of two
Skytraq Orion B16 receivers connected to one Taoglas
AP.25E.07.0054A active patch antenna each. The anten-
nas are mounted on opposite sides of the satellite to ensure
reception in all attitudes. The receiver was mainly chosen
because it was the most cost-effective option without alti-
tude or speed restrictions.
4.1.4 Sun Sensors
SOURCE uses a total of 16 OSRAM SFH-2430-
Z photodiodes as sun sensors. The low cost of those
COTS diodes and their flight heritage is what makes them
unique. In addition, the RAX mission showed that the
diodes feature a fast response time, high field of view and
high temperature operating range [13, 14]. Although a
half cone angle of 60◦ was specified by the manufacturer,
the RAX Mission showed that the cos α relationship of the
diodes used can achieve a half cone angle of up to 70◦ .
To protect the photodiodes from UV-degradation, Solar-
glas 0787 by Schott AG was chosen. The selected cover
glass is glued to the photodiodes with space qualified glue
DOW CORNING 93-500.
The SuS signals are routed through an OPA4196 amplifier
and a MAX1229 analog to digital converter. The corre-
sponding readout circuit is included on the housekeeping
boards.
4.1.5 Flight Heritage
The flown and planned flight heritage of the presented
COTS ADCS sensors is referenced in tab. 2. All sensors
are planned to fly also in other missions, while the SuS
already have done so.
4.2 Magnetic Actuators
To save cost, volume, power and mass, the attitude
control of SOURCE is provided by three in-house devel-
oped magnetorquers: one Air-core (AC) and two Cylin-
drical (C)-MGTs with ferromagnetic cores. Magnetor-
quers were chosen due to their high reliability [18, 19] and
IAC-22,E2,3-GTS.4,10,x70167
Table 2. Flight heritage of SOURCE COTS ADCS sen-
sors in other CubeSats
Heritage SuS MGM Gyro GNSS
Flown RAX-1 [14]
RAX-2 [14]
Planned GT-1 [15] PR-CuNaR2 PR-CuNaR2 PADDLES
[16] [16] [17] [antenna]
EIVE
low hardware complexity. All of the aforementioned as-
pects render them more convenient than other alternatives.
Unlike reaction wheels, for example, MGTs can actuate
autonomously as they do not need a momentum dump.
One downside of the MGT-only approach is the increased
complexity in attitude control algorithms, as torque can
only be applied perpendicular to the magnetic field vector
at any given time. Research in this field, however, is ad-
vancing fast [20] and this approach has already been suc-
cessfully demonstrated several times [21, 22]. This led to
various other missions to also adopt an MGT-only actua-
tion [23, 24].
Accordingly, the MGTs were developed, built and tested
by the team, as also has already been done in other mis-
sions [25, 26, 27, 28]. A unique advantage of this ap-
proach is that they could be individually adapted to the
satellite instead of vice versa, as would be the case with
a commercial MGT module. The configuration of one
AC-MGT and two C-MGTs was chosen as it combines
the best performance with a compact form factor. A soft
ferrite was selected core material for the C-MGTs. This
material is easily magnetized, but has low hysteresis char-
acteristics, meaning only a small remanent magnetic field
is produced by the ferromagnet after the electromagnet is
turned off.
4.2.1 Magnetorquer Development
The in-house development process of the MGTs has
already been presented in [29], and will only be outlined
briefly.
The optimal design is automatically derived by scripts,
which calculate all possible designs that fit the given con-
straints and propose the option with the highest magnetic
dipole moment. The major constraints are given by the
available space, power, available voltages and wire mate-
rials (electrical resistivity).
The manufacturing process was carried out using a wind-
ing machine. Secondary structures made from 3D-printed
plastic were used to guide the form of the coils during the
winding operation. To solidify the structure, space-grade
epoxy-resin was used.
The final step of the MGTs’ development is their func-
tional characterisation. The two common methods for
determining the magnetic dipole moment are either by
measuring the generated torque [30, 31] or the resulting
magnetic flux density [32]. Due to the availability of a
Page 5 of 14
 73rd International Astronautical Congress (IAC), Paris, France 18-22 September 2022.
Copyright 2022 by Mr. Nadim Maraqten. Published by the IAF, with permission and released to the IAF to publish in all forms.

highly accurate magnetometer (FGM3D) and for simplic- 4.3 Housekeeping Boards
ity reasons, the latter method was employed as proposed To accommodate the MGMs, gyroscopes, SuS read-
by Lee et al. [32]. This method allows for measurement outs and H-Bridge drivers for MGT control, SOURCE re-
of the dipole moment, hysteresis dipole moment and de- lies on two Housekeeping PCBs. These were designed,
cay time by measuring the magnetic flux density in on and soldered and qualified by the team. In order to minimize
off states of the MGTs at varying distances. The resulting cost, soldering the double-sided PCB was conducted man-
operational as well as the structural characteristics of the ually. Besides the aforementioned ADCS components,
Qualification Models (QMs) can be found in Table 3. the housekeeping boards contain thermal readout compo-
nents for the thermal subsystem and power conditioning
peripherals for payload applications.
Table 3. Operational and structural characteristics of the
QM air-core (AC) and cylindrical (C) magnetorquers
4.4 Hardware Cost Estimation
Characteristic AC C-67mm C-71mm
Outer Dimensions [mm³] 80*75*9 67*10*10 71*10*10 Estimating the true hardware cost of the SOURCE
Core Dimensions [mm²] 56*51 5*5 5*5 ADCS is difficult because of the number of prototypes and
Mass [g] 109 31 33 other iterations made during development. Instead, this
Power Consumption [W] 1.3 1.2 1.1 section assumes the standard model philosophy of one En-
Operating Voltage [V] 5 3.3 3.3 gineering Model (EM), QM and Flight Model (FM) each
Cu Wire diameter [mm] 0.355 0.3 0.3 to give a more accurate estimate of the resources needed
Windings 504 1515 1585
Dipole Moment [Am2 ] 0.47 0.44 0.49
to implement such a system for a future mission.
Hysteresis Dipole [Am2 ] 0 0.003 0.003 The estimated cost of the components is listed in Table 4.
Decay Time [ms] <4 <4 <5 The "No." column shows the count of this component in
one satellite times the number of model iterations. In the
case of the SuS cover glasses and MGTs, material could
The tests revealed, that the self-developed magnetor- only be purchased in bulk and therefore the total cost is
quers exceed the required magnetic dipole moments. It higher than needed for the number of models produced.
was found that the measured dipole moment of the AC- The miscellaneous item mainly includes periphery on the
MGT is in good agreement with the theoretical values, as Housekeeping Board (HKB) PCBs as well as some struc-
was also previously found in CubETH’s MGTs [31]. The tural parts and custom tooling. This puts the flight sys-
hysteresis and decay effects of the coils are both fairly
small and do not pose a problem for the ADCS. Table 4. ADCS Hardware Cost Overview
The three magnetorquers are integrated into one compact Component No. Unit FM Total
module. The AC is glued onto an empty Printed Cir- SuS 16 ∗ 3 1,25 C 20 C 60 C
cuit Board (PCB). The Cylinders are glued into small alu- SuS Glasses 16∗NA 6,25 C 100C 250C
minium interfaces which are placed on the other side of SuS Readout 1∗3 50 C 50 C 150C
the PCB and fixed onto the satellite’s main threaded rods. Gyro 1∗3 5C 10 C 30 C
The magnetorquers in their integrated form are shown in MGM 3∗3 23 C 69 C 207C
Fig. 5. They are controlled with motor drivers that are GNSS Rec. 2 ∗ 3 840C 1680C 5040C
connected to the Housekeeping Board. GNSS Ant. 2∗3 14 C 28 C 84 C
HKB PCBs 2 ∗ 3 180C 360C 1080C
MGT set of 3 1∗NA 25 C 25 C 270C
Misc. 1∗3 NA 400C 1200C
Total Hardware Cost: 2742C 8371C

Fig. 5. In-house developed air-core magnetorquer and available examples, the ISISSPACE iMTQ Magnetorquer
cylindrical magnetorquers (QMs) mounted on PCBs
tem price at 2742C and the total cost at 8371C, with the
two COTS GNSS receivers accounting for over 60% of
these figures. Direct comparisons with other commercial
systems are difficult due to differences in system perfor-
mance, specifications and dimensions, as well as the lack
of personnel costs or profit margins for university teams.
A clear cost benefit is still evident. To name two easily
Board alone costs 9750C [33] and the integrated Cube-
Space CubeADCS Y-Momentum system, with two MGTs
and a momentum wheel, is available for 22000 USD [34].

IAC-22,E2,3-GTS.4,10,x70167 Page 6 of 14
 73rd International Astronautical Congress (IAC), Paris, France 18-22 September 2022.
Copyright 2022 by Mr. Nadim Maraqten. Published by the IAF, with permission and released to the IAF to publish in all forms.
Neither product includes a GNSS receiver. This shows
that the chosen ADCS design is a very cost-effective op-
tion for university teams, while still enabling reliability
through redundancy and rigorous testing, as shown in Sec-
tion 5.
5. Hardware Testing
Extensive hardware testing is required to qualify the
system and especially to approve the usage of the var-
ious COTS components and the self-built MGTs. The
tests were carried out taking into account the low-cost
strategy without compromising their soundness and can
be divided into four major groups: functional, thermal-
vacuum, shaker and electromagnetic compatibility tests.
It is known that radiation is an issue which demands for
consideration, especially in long term space missions and
those with designs including COTS and self-built compo-
nents. The SOURCE mission, however, is a rather short-
term mission (≤ 2 years), planned to fly in Low Earth Or-
bit (LEO) within the protection of Earth’s magnetosphere.
A double and triple redundancy philosophy for critical
ADCS components was chosen. Moreover, radiation tests
can be very expensive and complex. As a consequence of
all of these aspects, it was decided not to conduct further
tests to distinctly evaluate the system’s response to radia-
tion.
5.1 Functional Testing
The functionality of all components is extensively
tested in the configuration planned for the satellite, taking 6. Attitude Determination and Control Laws
into account the peculiarities (e.g. hysteresis of MGTs) of
the different components. Performance qualification tests
are also conducted for accurate determination of the capa-
bilities and characteristics of the components.
5.2 Thermal-Vacuum Testing
To verify the components’ resistance against the harsh
conditions in space, they were probed in a Thermal Vac-
uum Chamber (TVac). These tests were conducted with
special care to avoid running the risk of repeating them.
Accordingly, conservative temperature ranges for the ther-
mal cycling were chosen. Instead of determining the
ranges by adding margins to the simulation results, it was
decided to use the maximum rated temperatures for the
respective components.
To minimize the number of tests, component tests with
similar temperature ranges were combined while keep-
ing the differences in thermal conduction between the
components in mind. This can lead to different temper-
ature plateaus and failure of the verification if the thermal actuators.
ranges are not chosen generously enough. A second and
possibly more important reason for choosing very conser- 6.2 Guidance
vative temperature ranges is the risk of errors in the sim-
ulation. If the thermal simulation results are wrong at the in sun pointing attitude. In Idle Mode SOURCE points the
IAC-22,E2,3-GTS.4,10,x70167
time of testing and are corrected afterwards, there is a risk
that the qualification of the components must be repeated
if a sufficiently large uncertainty has not been factored in.
5.3 Shaker Testing
To verify the structural integrity during launch, shaker
tests were conducted according to European Coopera-
tion for Space Standardization (ECSS) standards. While
smaller CubeSats tend to perform shaker tests on flight
hardware, the SOURCE CubeSat was probed using a
separate structural QM. This structural model contains
all component QMs as well as dummies for the COTS
components. To imitate launch conditions, a designated
shaker cage was constructed to simulate the behaviour ex-
pected from the CubeSat deployer.
5.4 Electromagnetic Compatibility Testing
To prevent the individual components from interfer-
ing with each other, the electromagnetic compatibility is
tested. This is especially important in the case at hand
since self-built and COTS components are used which are
not tuned for compatibility with any other component. In-
terference frequencies emitted by the individual compo-
nents are measured in order to install filters or shielding
if necessary. If all components are controlled at the same
time, problems in the lines and in the power supply can
also occur. However, these can be detected and corrected
by testing.
Following the standard algorithm framework,
SOURCE’s ADCS is composed of guidance, naviga-
tion and control algorithms. The spacecrafts ADCS
logic is tested in a dynamic 6-DoF simulation environ-
ment named GAFE. The subsystem utilizes the sensors
in Section 4.1 according to the system modes in Section 3.
6.1 Simulation environment
GAFE is a design framework for satellites developed
by the ESA technology department ESTEC and part-
ners during the ESA-GSTP study. The ADCS team of
SOURCE mainly works with the time domain 6-DoF sim-
ulation part of GAFE. While developing and testing Fail-
ure Detection, Isolation, and Recovery (FDIR) techniques
is the main use-case of this simulation environment, it can
also be a powerful tool when testing attitude determina-
tion and control algorithms and the usage of sensors and
Most of the mission’s payload activities are executed
Page 7 of 14
 73rd International Astronautical Congress (IAC), Paris, France 18-22 September 2022.
Copyright 2022 by Mr. Nadim Maraqten. Published by the IAF, with permission and released to the IAF to publish in all forms.

normal of the solar arrays sref (positive X axis of the ge- QUEST
ometry frame, see Fig. 1) along the measured sun vector
in the geometry frame s. In Safe Mode the reference di- Payload Camera
rection is defined as the satellite’s principal axis of inertia Gyro SUS MGM Clock GPS STR
M
to further stabilize the rotation of the satellite around the
reference direction. Consequently, the satellite aligns the Brightness
Time Position
normal of the solar arrays to the sun at a slight angle. The
target quaternion is switched from sun pointing to an iner- Sun Vector Sun Vector Magnetic
tial Earth direction when required for meteor observation Calculation Model Field Model

campaigns. bB rB

bS

rS
Attitude
6.3 Navigation Angular Velocity Quaternion
Kalman Filter
SOURCE’s navigation system is in charge of deter-
mining the satellite’s position and attitude continuously
Quaternion
Initial Attitude Quaternion
from commissioning to burning up in the atmosphere.
Real Time
6.3.1 Sun Vector Determination Attitude

As mentioned in Section 4.1 the SOURCE mission

is equipped with a total of 16 SuS. The photodiodes are
mounted in redundant pairs on each surface of the satel-
lite except for the +X axis of the geometry frame (see Fig.
1). The positive X axis aligns with the normal of the solar
arrays.
A pyramid geometry mount for the SuS was installed on
the face of the solar arrays to enhance sun vector determi-
nation accuracy in case of a low sun pointing error. Spe-
cific details regarding the implementation of the SuS and
the computation of the sun vector can be taken from [35].
6.3.2 Multiplicative Extended Kalman Filter
For attitude determination, a Multiplicative Extended
Kalman Filter (MEKF) with a kinematic propagation of
the satellite attitude is chosen. Fig. 6 depicts the input
states of the MEKF. The output states are the satellite at-
titude quaternion q0 and the bias of the gyroscope β0 . The
chosen MEKF is based on the filter, presented in [1]. For
the exact derivation of the SOURCE Kalman filter, please
refer to [36].
By initializing the MEKF with an inaccurate state,
the assumption of linearization of the filter is not valid.
This could decrease the performance in the steady state or
could even lead to a divergence of the attitude estimation.
To overcome this issue, the Quaternion Estimator
(QUEST) algorithm is used to compute the initial condi-
tions. It was shown that the chosen filter is able to achieve
a sufficient determination accuracy to fulfil all mission re-
quirements during sun phase (see Section 2) [36].
Fig. 6. Attitude Determination Scheme
As this method will be flight-tested during the SOURCE
mission, it will not be used as a primary attitude determi-
nation sensor. However, the MeSHCam can be integrated
into the attitude determination chain, whenever a reliable
measurement is available.
6.5 Control
Control over SOURCE’s attitude is achieved using an
MGT-only approach with three MGTs, one along each
axis. As discussed in Sec. 4.2, this increases control
complexity as torque can only be applied perpendicular
to Earth’s magnetic field, but is more convenient in terms
of cost, mass, power and volume.
6.5.1 Detumble Mode Control Law
In Detumble Mode, the only active sensor is the mag-
netometer. Its measurements are utilized to estimate the
portion of the angular velocity which is perpendicular to
the magnetic field ω⊥B . However, due to fluctuations in
the angular velocity induced by nutation perpendicular to
the magnetic field, it was shown not to be sufficient to
monotonously reduce the angular velocity of the satel-
lite. Therefore, the Detumble algorithm follows the ap-
proach of monotonously reducing the rotational energy of
SOURCE. This approach has been shown to be sufficient
for use. The Proportional B-Dot-Law was derived as

Ḃ
6.4 Experimental Use of a Payload Camera as a Star m = −k · 2 . (1)
Tracker |B|
The payload camera MeSHCam, will be used as an ex- The gain parameter k has been optimized specifically
perimental star tracker. The camera’s images are fed into a for SOURCE, utilizing the MATLAB based tool GAFE
open-source star tracking algorithm [37]. This algorithm described in Section 6.1. Other approaches for rotational
is able to directly extract the current attitude quaternion. energy minimization have been considered, which reduce

IAC-22,E2,3-GTS.4,10,x70167 Page 8 of 14
 73rd International Astronautical Congress (IAC), Paris, France 18-22 September 2022.
Copyright 2022 by Mr. Nadim Maraqten. Published by the IAF, with permission and released to the IAF to publish in all forms.

the initial angular velocity in a comparable time span to
even lower values. However, to further enhance mis-
sion robustness, the Proportional B-Dot-Law has been se-
lected since it does not depend on the inertia tensor of the
satellite. A comprehensive analysis and the derivation of
the analysed control laws for the Detumble Mode can be
found in [38].
the Indirect Lyapunov Method [38]. The investigation
found Lyapunov-stability for the controller with an ex-
ception. If the alignment torque Mal has a different sign
than the rate error parallel to the magnetic field vector
(∆∥B = (ω −ωref )∥B ) stability is not guaranteed. Mani-
festation of the conditional instability is represented in the
following excerpt from the Lyapunov function

6.5.2 Safe Mode and Idle Mode Control Law

δ
In 2017 the IRS launched the small satellite mission V̇3 = −kal · · ∆∥B ◦ (sref × s). (5)
sin(δ)
Flying Laptop (FLP) which relies on a PD-type controller
for its Safe Mode attitude control. The FLP Safe Mode With V̇3 being a summand of the Lyapunov function
control law was designed for MGT-only operation and has V̇ in [38]. To ensure stability [38] suggests compensation
been adapted for all control modes of SOURCE, including in the form of a short term adaption of gains to meet the
Safe Mode, Idle Mode and Pointing Mode. The adapta- following inequality:
tion of the FLP Safe Mode controller to the higher oper-
ational modes of the SOURCE mission is possible since
a moderate pointing accuracy in higher modes is required ! δ ∆∥B ◦ (sref × s)
kω > −kal · · . (6)
(see Section 2). The adaptation is nonetheless demand- sin(δ) ∆ ◦ ∆⊥B
ing, considering the challenge of under-actuation due to
the sole use of MGTs. The compensation momentarily increases the rate
Although the control law remains unchanged during the damping moment Mω , which conserves the fix point but
SOURCE operational modes, there is an adaption consid- may lead to an increased attitude error. Verification of
ering the number of sensors used for the attitude determi- the control law and tuning of the controller gains kω and
nation outlined in Table 1. Additionally, the Idle Mode kal are accomplished through simulation runs in GAFE
and Pointing Mode controller will be active during the (see 6.1). This dynamic 6-DoF simulation environment
eclipse phases while the Safe Mode controller is only ac- allows for modelling of the sensors listed in Section 4.1.
tive during sun phases to conserve the battery charge. A Additionally, actuators and the following environmental
third step is the adaption of gain parameters to the specific perturbations are included for SOURCE:
mode.
Fundamentally the control law is laid out as a non-linear • Atmospheric forces and torques (Harris-Priester den-
PD-type controller. Linearization has been deemed im- sity model and aerodynamic drag model as described
practical considering that the attitude control system of in [11])
SOURCE is required to cope with alignment errors of any • Gravity Gradient torque (JGM-3)
magnitude. Therefore defining a linearization state is not
possible. The control law, shown in (2), is constructed • Solar Radiation forces and torques (SolarFluxDen-
of an alignment term Mal and a rate damping term Mω sity model developed for [38])
adding up to the commanded torque
• Torques caused by Residual Dipole (World Magnetic
M = Mal + Mω . (2) Model)

The alignment torque in equation (3) is proportional to The goal of the parameter optimization is to tune the
the angular error between the sun vector and the reference controller gains kω and kal of the PD-type controller to
direction of SOURCE. It amounts to the expected orbital environment. The process is ideally
pursued in a dynamic 6-DoF simulation as provided by
δ
Mal = kal · (sref × s) · . (3) GAFE, considering that MGTs rely on a time dependent
sin(δ) magnetic field vector. For the gain optimization, the sim-
ulation is run on a coarse grid of gain pairs first. A grid of
While the rate damping torque amounts to
gain pairs is defined as an array that provides every possi-
ble combination of the gains kω and kal within specified
Mω = −Kω · (ω − ωref ), (4)
limits. Then the controller performance is evaluated based
where Kω is constructed as a diagonal 3 × 3 ma- on a simulation for each gain pair. Among the evaluation
trix with identical entries kω on its principal diagonal. criteria are stability, pointing accuracy, rate damping and
The alignment gain kal is of scalar dimension. The sta- actuator saturation. In an iterative approach, the simula-
bility of the PD-type control law was investigated using tion is repeated on a much denser grid of gain pairs in

IAC-22,E2,3-GTS.4,10,x70167 Page 9 of 14
 73rd International Astronautical Congress (IAC), Paris, France 18-22 September 2022.
Copyright 2022 by Mr. Nadim Maraqten. Published by the IAF, with permission and released to the IAF to publish in all forms.

XB
the immediate vicinity of a promising pair of gains. In-
creasing the resolution of gain values in an iterative fash- ZA

ion leads closer to the optimized gain parameter value XA ZB

with each iteration. Further investigation will analyse the α
feasibility of an additional three-axis stabilization for the β
Pointing Mode of the SOURCE mission.

7. Analysis of the ADCS Performance during Re-

entry
In order to ensure SOURCE’s functionality in the YB

early re-entry phase and to meet the scientific goals vrel YA

of the mission, predictions of the CubeSat’s aero-
thermodynamic behaviour at altitudes below 200 km were
Fig. 7. SOURCE’s body frame and the aerodynamic ref-
performed [39]. For this, reference Direct Simulation
erence frame that follows the relative velocity vector.
Monte Carlo (DSMC) flow simulations were carried out
that yielded results on the heat flux and pressure distribu-
tions around SOURCE during re-entry [4]. This not only date and simulated trajectories [39]. To enable predictions
enabled sensor selection and estimations on the minimal about the dynamic behaviour of SOURCE, DSMC simu-
operational altitude of the satellite but laid the founda- lations at various angles of incidence were carried out at
tion for further analysis of the CubeSat’s aerodynamic be- an altitude of 150 km to determine aerodynamic coeffi-
haviour through dynamic 6-DoF simulations. Since aero- cients as required for 6-DoF simulations. The pressure
dynamic disturbances are expected to exceed the ability distribution around the satellite’s body for one specific
of the ADCS to counteract at altitudes below ∼240 km, angle of attack is shown in Fig. 8. Integration of this
all control over the satellite’s attitude is impaired below
this threshold. Although there are no requirements on the
ADCS for re-entry due to this reason, best-possible atti-
tude determination provides valuable insight when eval-
uating the pressure, heat flux and FIPEX sensor data
and is therefore pursued. Predicting or even controlling
SOURCE’s attitude during this final mission phase is even
more valuable for the mission and was therefore investi-
gated further.

7.1 Geometry and Coordinate System

For the dynamic 6-DoF simulations, boundary condi- Fig. 8. Surface pressure distribution around SOURCE for
tions and aerodynamic parameters are derived from spec- an altitude of 150 km an angle of attack of α = 60° and
ified geometry and coordinate systems. A simplified sideslip angle of β = 0°, from DSMC simulation results.
model of the SOURCE CubeSat consisting of a cuboid
core, solar panels and an additional front facing cylinder pressure distribution over all surfaces yields the net forces
as shown in Fig. 2 is used for the DSMC simulations. The and torques that act on SOURCE for one particular angle
following analysis uses an aerodynamic reference system of incident flow. Using the static drag equations, aerody-
that is defined using the relative satellite velocity in the namic coefficients can be derived and used for stability
B
body frame vrel and the X-axis of the satellite XB B: analyses:
B
vrel  
ZB
A =− B ∥
, B
YA = ZB B
A ×XB , XB B B
A = YA ×ZA . CL
1
∥vrel CF =  CS  = FA (8)
ρ B ∥2 drag
(7)
CD 2 Aref ∥vrel
Both coordinate systems together with the angle of attack
α and sideslip angle β can be viewed in Fig. 7.  
CM X
1
7.2 DSMC Simulation Results CM ρ =  CM Y  = B 2
MB drag . (9)
2 Aref Lref ∥vrel ∥
CM Z
For the DSMC pressure and heat flux simulations, ex-
treme environmental conditions have been defined based Hereby, input variables of the numerical simulations
on the state of the atmosphere at the expected re-entry are used in the denominator. The reference area of

IAC-22,E2,3-GTS.4,10,x70167 Page 10 of 14
 73rd International Astronautical Congress (IAC), Paris, France 18-22 September 2022.
Copyright 2022 by Mr. Nadim Maraqten. Published by the IAF, with permission and released to the IAF to publish in all forms.
Aref = 0.075 m2 and length of Lref = 0.363 m were
geometrically derived. The thermospheric mass density
was taken from the NRLMSISE-00 atmospheric model
with a value of ρ = 2.595 × 10−9 kg m−3 for the used
altitude of 150 km [40]. The relative speed of the molec-
ular flow in the spacecraft’s body coordinates was set to
B
∥vrel ∥ = 7.891 km s−1 . While the coefficients for forces
are calculated in the aerodynamic system using the force
vector FA B
drag , the torques Mdrag apply to SOURCE’s
principal body axes. The so attained coefficients can be
validated by comparing them to satellites that share a sim-
ilar geometry with SOURCE [39].
Rearrangement of Equations (8) and (9) offers the abil-
ity to calculate forces and torques on SOURCE for other
atmospheric densities and velocities, thus enabling the
modelling of aerodynamic disturbances in dynamic sim-
ulations. The derived coefficients only exist for selected
angles of incident flow (α, β) and need to be interpolated
for interim values to be used in dynamic simulations [11].
Biharmonic spline interpolation is used to estimate all in-
terim grid values. This creates the meshed surfaces in
three dimensional space that are depicted in Fig. 9. To
avoid time-consuming calculations for every time step,
pre-constructed lookup tables are used that can vary in
resolution depending on the desired accuracy.
7.3 Dynamic Simulation Results
The attained torque coefficients for the three princi-
pal body axes were used for preliminary analysis and
promised a statically stable attitude of SOURCE at
α = 90° (the +X-axis of SOURCE coincides with the
airstream vector) [39]. Whether SOURCE is also pas-
sively stable in the dynamic environment of re-entry was
investigated in 6-DoF simulations with GAFE. Further-
more, the performance of the Detumble controller de-
scribed in Section 6.5.1 was compared to a baseline of
no active control (see Fig. 10). When rate dampening
is active, angular velocities are kept to a minimum with
1.5 ° s−1 only exceeded after 12 h (below 130 km). This
enables accurate attitude determination until the battery is
depleted or the thermal limits are reached, since the algo-
rithm was tested to work reliably below 10 ° s−1 .
The angle of incident flow, also depicted in Fig. 10,
promises a steady state of SOURCE around two axes. The
oscillation of the sideslip angle β can be subjected to the
discontinuity of the Euler angle representation at this at-
titude and is not an attribute of instability. The above de-
scribed statically stable orientation of SOURCE at a pitch
angle of α = 90°, together with the dampening character-
istics of the Detumble controller, achieve a dynamically
stable attitude. This dynamic stability cannot be observed
for the case of inactive ADCS. This ability may prove very
useful when construing strategies for the operations dur-
ing the final mission phase.
IAC-22,E2,3-GTS.4,10,x70167
Furthermore, the orbit decay in the stable case is in-
creased, resulting in a shorter lifetime, as the drag coef-
ficient is highest for the α = 90° orientation (see Fig. 9).
Although active attitude control during re-entry cuts the
orbital lifetime in half, the potential advantages of lower
angular velocities and thus better acquisition of scientific
data may be more desirable [11]. A final validation of
the entire aero-thermodynamic simulation procedure —
beginning with the free molecular flow DSMC simula-
tions up until the coefficient interpolation and the dynamic
modelling using GAFE — can be performed by reviewing
flight data. For this, heat flux measurements and the atti-
tude estimation, that is derived from sun sensor, magne-
tometer and gyro data, will be compared to the simulated
descents.
8. Conclusions
To further increase space accessibility, it is imperative
to decrease the associated costs. This paper gives insight
into the low-cost ADCS of the student-led 3U+ CubeSat
SOURCE, which will fly in LEO and needs its ADCS
especially for sun-pointing of the solar panels and pre-
cise attitude determination for the scientific payloads. By
exploitation of two main strategies it was possible to de-
velop a reliable ADCS at considerably low cost. These
are the use of COTS components and in-house devel-
opment wherever possible. With the SOURCE mission
those components will gain flight heritage making them
attractive for use in other missions.
The attitude determination is based on magnetometers,
sun sensors, gyroscopes and a GNSS receiver. Attitude
control is achieved by three self-built magnetorquers. All
of these components are subject to extensive functional,
thermal-vacuum, shaker and electromagnetic compatibil-
ity testing to affirm their reliability. Guidance, control
and navigation algorithms are tested in GAFE, a dynamic
6 DoF simulation environment developed by ESA. The
ADCS performance during re-entry is of upmost impor-
tance for the scientific mission of the satellite and was
therefore investigated with special care. It was found that
achieving a stable attitude of SOURCE during early re-
entry is possible with the use of the rate dampening De-
tumble Mode. Whether this strategy is pursued will be
investigated further.
The hardware cost of the flight system was estimated to
2742 C and that of the overall system to 8371 C, with the
two COTS GNSS receivers accounting for over 60% of
these figures. The developed ADCS is significantly more
cost-effective than other commercially available systems,
which is especially interesting for other educational and
low-cost missions with limited budgets. Adopting the pre-
sented strategies and architectural aspects of the system
will provide an opportunity for cost reduction and might,
if successfully implemented, even be a mission enabler.
Page 11 of 14
 73rd International Astronautical Congress (IAC), Paris, France 18-22 September 2022.
Copyright 2022 by Mr. Nadim Maraqten. Published by the IAF, with permission and released to the IAF to publish in all forms.

(a) Lift coefficient CL (b) Side force coefficient CS (c) Drag coefficient CD

(d) Roll torque coefficient CM X (e) Pitch torque coefficient CM Y (f) Yaw torque coefficient CM Z

Fig. 9. Visual representation of the surfaces generated by the aerodynamic coefficient interpolation for SOURCE,
circles mark the data points from DSMC simulations and boundary conditions.

(a) No attitude control (b) Active rate dampening

Fig. 10. Angular velocities and angle of incident flow of uncontrolled and rate dampened spacecraft starting at an
altitude of 200 km.

IAC-22,E2,3-GTS.4,10,x70167 Page 12 of 14
 73rd International Astronautical Congress (IAC), Paris, France 18-22 September 2022.
Copyright 2022 by Mr. Nadim Maraqten. Published by the IAF, with permission and released to the IAF to publish in all forms.
Acknowledgements [11]
The authors would like to express their deep gratitude
to the SOURCE project leaders, Michael Lengowski and
Daniel Galla, for doing their best in making this unique
hands-on experience of building a satellite already as a [12]
student possible. Moreover, the team is sincerely thankful
to the ESA - Fly Your Satellite! team, who gave indespen-
sible advice for the prosperity of this mission.
[13]
References
[1] F. L. Markley and J. L. Crassidis. Fundamentals
of spacecraft attitude determination and control,
Springer, 2014.
[2] E. Messerschmid and S. Fasoulas. Raumfahrtsys- [14]
teme, vol. 2. Springer, 2011.
[3] D. Galla, S. Klinkner, G. Herdrich, et al. The Ed-
ucational Platform SOURCE - A CubeSat Mis-
sion on Demise Investigation Using In-Situ Heat [15]
Flux Measurements, 70th International Astronauti-
cal Congress. 2019.
[4] D. Galla, S. Klinkner, G. Herdrich, et al. The De-
tailed Design of the SOURCE Satellite for Demise
Investigation, 33rd International Symposium on
Space Technology and Science. 2022.
[5] R. Schweigert, A. Stier, M. Lengowski, et al. Phase [16]
A Study for the Earth Observation and Technology
Demonstration CubeSat SOURCE, 12th IAA Sym-
posium on Small Satellites for Earth Observation. [17]
2019.
[6] T. Stäbler, K. Chen, T. Cziep, et al. IRAS-New
technologies for low cost satellites, 72nd Interna-
tional Astronautical Congress. 2021. [18]
[7] A. Stier, R. Schweigert, D. Galla, et al. Combina-
tion of Interdisciplinary Training in Space Technol-
ogy with Project-Related Work through the Cube-
Sat SOURCE, 3rd Symposium on Space Educa- [19]
tional Activities. 2019. DOI: 10.29311/2020.
46.
[8] H. Jones. The recent large reduction in space
launch cost, 48th International Conference on En-
vironmental Systems. 2018. [20]
[9] M. Ehresmann, F. Hild, K. Grunwald, et al. PA-
PELL: Mechanic-free Actuators through Ferroflu-
ids, 12th IAA Symposium on Small Satellites for
Earth. 2019.
[10] M. von Ehrenfried. Ingenuity, Perseverance and [21]
the Mars 2020 Mission: Follow the Science to
Jezero Crater. Springer International Publishing, [22]
2022, pp. 111–125. ISBN: 978-3-030-92118-7.
IAC-22,E2,3-GTS.4,10,x70167
P. Haufe. Aerodynamic Simulation of the
SOURCE CubeSat in the Early Re-entry Phase,
Bachelor’s Thesis. Institute of Space Systems,
University of Stuttgart, 2021.
Bosch. IMU: BMI270. URL: https : / / www .
bosch - sensortec . com / products /
motion-sensors/imus/bmi270 (visited on
08/18/2022).
J. C. Springmann, A. J. Sloboda, A. T. Klesh, et
al. The attitude determination system of the RAX
satellite, Acta Astronautica 75 (2012), pp. 120–
135. DOI: 10 . 1016 / j . actaastro . 2012 .
02.001.
J. C. Springmann and J. W. Cutler. Flight results
of a low-cost attitude determination system, Acta
Astronautica 99 (2014), pp. 201–214. DOI: 10 .
1016/j.actaastro.2014.02.026.
N. G. Guecha-Ahumada and E. G. Lightsey. De-
sign of the Attitude Estimation Algorithms for the
GT-1 CubeSat, Georgia Institute of Technology.
2020. URL: https : / / www . ssdl . gatech .
edu / sites / default / files / ssdl -
files / papers / mastersProjects /
GuechaG-8900.pdf (visited on 08/08/2022).
I. Narrett and B. Strabel. Magnetic Field Measure-
ment Suite for CubeSat Applications, American
Geophysical Union Fall Meeting. 2020.
D. Guglielmo and R. Bevilacqua. Propellant-
less atmospheric differential drag LEo spacecraft
(PADDLES) mission, Proceedings of the 2014
SmallSat Conference. 2014.
R. Wisniewski and J. Stoustrup. Periodic H2 syn-
thesis for spacecraft attitude control with magne-
torquers, Journal of Guidance, Control, and Dy-
namics 27.5 (2004), pp. 874–881.
P. Wang and Y. B. Shtessel. Satellite attitude
control using only magnetorquers, Proceedings
of the 1998 American Control Conference. ACC
(IEEE Cat. No. 98CH36207). Vol. 1. IEEE. 1998,
pp. 222–226.
M. Ovchinnikov and D. Roldugin. A survey on
active magnetic attitude control algorithms for
small satellites, Progress in Aerospace Sciences
109 (2019). DOI: 10 . 1016 / j . paerosci .
2019.05.006.
P. Hoffmeyer. The Ørsted satellite project, Air &
Space Europe 2.5 (2000), pp. 74–79.
S. Rückerl, D. Messmann, N. Appel, et al. First
Flight Results of the MOVE-II Satellite, Proceed-
ings of the Small Satellite Conference. 2019.
Page 13 of 14
 73rd International Astronautical Congress (IAC), Paris, France 18-22 September 2022.
Copyright 2022 by Mr. Nadim Maraqten. Published by the IAF, with permission and released to the IAF to publish in all forms.

[23] N. Jovanovic, B. Riwanto, P. Niemelä, et al. De- [35] N. Pentke, J. Schaefer, and T. Karl-Heinz Schwarz.
sign of Magnetorquer-Based Attitude Control Sub- Comparison of the Low-Cost Sun Sensors of the
system for FORESAIL-1 Satellite, IEEE Journal SOURCE and EIVE CubeSats, 36th Annual Small
on Miniaturization for Air and Space Systems 2.4 Satellite Conference.
(2021), pp. 220–235. [36] C. Luckas. Entwicklung und Verifikation von
[24] Z. Hu, T. Timmons, L. Stamat, et al. Development einer Lagebestimmung des studentischen CubeSats
of a 10g femto-satellite with active attitude control, Source, Bachelor’s Thesis. Institute of Space Sys-
17th Reinventing Space Conference. 2019. tems, University of Stuttgart, 2020.
[25] D. Miller. Design optimization of the CADRE [37] Astrometry. Documentation. URL: https :
Magnetorquers. University of Michigan, Ann Ar- / / astrometry . net / doc/ (visited on
bor, May 2, 2013. URL: https : / / www . 08/08/2022).
aerospades . com / uploads / 3 / 7 / 3 / 2 / [38] F. Tuttas. Design of an Attitude Control System for
37325123 / cadre _ torquers . pdf (visited the Nanosatellite Student-Operated University Re-
on 08/08/2022). search Cubesat for Education (SOURCE), Bache-
[26] S. Ahmed Khan, A. Ali, Y. Shiyou, et al. Optimized lor’s Thesis. Institute of Space Systems, University
Design and Analysis of Printed Magnetorquer for of Stuttgart, 2018.
a 3-U Nano-Satellite, Journal of Aerospace Engi- [39] D. Galla, H. Fischer, and P. Haufe. In-Situ Mea-
neering 35.1 (2022), p. 04021103. surements in Early Phase Re-entry of the SOURCE
[27] G. Manzoni, Y. L. Brama, and M. Zhang. CubeSat for Numerical Simulation Validation,
Athenoxat-1, Night Vision Experiments in LEO, 2nd International Conference on Flight Vehicles,
Proceedings of the Small Satellite Conference. Aerothermodynamics and Re-entry Missions &
2016. Engineering (FAR). Heilbronn, Germany, 2022.
[28] V. P. Katkoori et al. Simulation and Selection of [40] US Naval Research Laboratory. NRLMSISE-00 At-
Detumbling Algorithms for 3U CubeSat, Interna- mosphere Model. URL: https : / / kauai .
tional Astronautical Congress. 2019. ccmc . gsfc . nasa . gov / instantrun /
[29] N. Maraqten, P. Haufe, C. Vogt, et al. A Guide msis (visited on 08/08/2022).
to Self-Built Low-Cost Magnetorquers as will be
used in the 3U+ CubeSat SOURCE, 12th European
Cubesat Symposium. 2021.
[30] I. Indir, K. Sever, I. Vnučec, et al. Design and Op-
timization of Air Core Magnetorquers for Attitude
Control of LEO Nanosatellites, 2021 International
Symposium ELMAR. IEEE. 2021, pp. 135–138.
[31] S. Rossi, A. Ivanov, G. Faure, et al. Cubeth adcs
design, implementation and validation tests, 66th
International Astronautical Congress, IAC, Jaruse-
lam, Israel. 2015.
[32] J. Lee, A. Ng, and R. Jobanputra. On determin-
ing dipole moments of a magnetic torquer rod-
experiments and discussions, Canadian aeronautics
and space journal 48.1 (2002), pp. 61–67.
[33] ISISSPACE. Magnetorquer Board (iMTQ). URL:
https : / / www . cubesatshop . com /
product / isis - magnetorquer - board/
(visited on 08/05/2022).
[34] CubeSpace Satellite Systems. Closed Loop ADCS.
URL : https : / / www . cubesatshop . com /
wp - content / uploads / 2020 / 03 /
CubeADCS-V1_2.pdf (visited on 08/05/2022).

IAC-22,E2,3-GTS.4,10,x70167 Page 14 of 14

View publication stats