SYSTEM DESIGN AND FABRICATION FOR MI CROSATELLITE RELATIVE

NAVIGATION EXPERIMENT

A Thesis

by

ANDREW LEWIS TUCKER

Submitted to the Office of Graduate and Professional Studies of

Texas A&M University
in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

Chair of Committee, Helen L. Reed

Committee Members, Daniele Mortari
Darren DePoy

Head of Department, Srinivas Rao Vadali

May 2021

Major Subject: Aerospace Engineering

Copyright 2021 Andrew Tucker

 ABSTRACT

Satellites are a critical element of the modern world, and designers continue to

increase their capability while significantly reducing their size, which has put space

missions within the reach of Universities. Microsatellites in the 10-100 kg size class are

now able to perform a sizable amount of tasks in a relatively small and inexpensive

package. Texas A&M University's second foray into space featured a 50 kg

microsatellite designed and manufactured by students within the AggieSat Lab Student

Satellite Program.

AggieSat4 was the second satellite fielded by AggieSat Lab under the NASA

Low-earth Orbiting Navigation Experiment for Spacecraft Testing Autonomous

Rendezvous and docking (LONESTAR) campaign. The LONESTAR campaign's goal

was to partner design labs from Texas A&M and the University of Texas at Austin to

build pairs of satellites to perform navigation experiments. A series of four missions

would culminate with the two paired spacecraft performing autonomous rendezvous and

docking.

AggieSat4 was designed and fabricated from 2010 to 2015, delivered to the

International Space Station in December 2015, and released into low Earth orbit in

January 2016. During this process a great deal of knowledge was gained by the students

as to how to design a spacecraft mission to meet a set of requirements, how to design

and engineer a spacecraft to carry out this mission, and how to fabricate and assemble
ii
the spacecraft as designed. Many tips, tricks, and lessons from hindsight were learned

along the way.

The requirements and mission concept of operations development for AggieSat4

will be presented, along with the engineering design process, resulting configuration,

fabrication process, and some of the tips, tricks, and lessons learned. These topics can

serve as a starting guide for students and others designing their own space missions, with

the goal of helping them identify the processes and items of consideration to help meet

their mission requirements.

iii
 ACKNOWLEDGEMENTS

I would like to thank Dr. Mortari and Dr. DePoy for sitting on my committee, as

well as for their guidance through coursework. I would also like to thank Gail Rowe for

helping me navigate all the administrative aspects of the degree process.

I would like to thank my parents and family for providing the resources and

preparation for my education, as well as the never ending love and support.

Thanks to Mr. David Kanipe and Mr. Darryl May of NASA's Aeroscience and

Flight Mechanics Division and Johnson Space Center for starting and managing the

LONESTAR program. Thanks to Mr. Joseph Perez, former AggieSat Lab director for

teaching me about hardware development and spaceflight integration. Thanks to Mr.

John Graves, AggieSat2 Program Manager and AggieSat Lab Manager for his leadership

and for teaching me about spacecraft operations, as well as developing radio hardware

for AggieSat4. And thanks to Mr. Zane Singleton for providing the technical interface

between AggieSat Lab and NASA.

A great deal of gratitude goes to my friends, co-workers, and classmates who

worked at AggieSat Lab on both AggieSat2 and AggieSat4. Without your hard work

and dedication, we would never have been able to make these missions happen. We

made a great team!

Thanks to everyone who participated at the NASA Johnson Space Center, the

United States Department of Defense Space Test Program, MEI Technologies,

Oceaneering Space Systems, our colleagues at the University of Texas at Austin, the

iv
University of Houston, ISS astronauts Scott Kelly and Tim Peake, and every amateur

radio operator worldwide who made this program possible end-to-end.

Thanks to groups that contributed funds and support to AggieSat4, to the Edward

"Pete" Aldridge '60 Professorship, Sinclair Interplanetary, VectorNav, The Aerospace

Corporation, and Lockheed Martin for their support. The team is also grateful to Boeing,

SpreadSheetWorld, Microhard Systems Inc., Texas Space Grant Consortium, National

Science Foundation Research Experiences for Undergraduates Program/Undergraduate

Summer Research Grants Program (at Texas A&M), and Sandstorm A&M Mother’s

Club.

AggieSat4 would not have been possible without the dedication and professional

support from Ms. Rebecca Marianno, Ms. Colleen Leatherman, Dr. Kristi Shryock, Mr.

Rodney Inmon, Mrs. Yolanda Veals, Mrs. Andrea Loggins, the Physics and Astronomy

Machine Shop, Dr. Rodney Bowersox (as Aerospace Engineering Department Head),

Mr. David Breeding (Engineering Safety Office), Dr. Gregory Huff and the

Electromagnetic and Microwave Laboratory team, and The Battalion.

And most importantly I would like to thank my committee chair Dr. Reed. I

can't begin to state just how much support, encouragement, resources, and time she has

freely given to me which have shaped me into the person I have become. I have been

truly blessed to have had her as a mentor and friend, and any student should be so

fortunate to have a teacher like her!

v
 CONTRIBUTORS AND FUNDING SOURCES

Contributors

This work was supervised by a thesis committee consisting of Professor Helen

Reed and Daniele Mortari of the Department of Aerospace Engineering. and Professor

Darren DePoy of the Department of Physics and Astronomy.

Funding Sources

Graduate study was supported by a Graduate Assistant in Research position from

Texas A&M University.

This work was also made possible in part by a Grant (NNJ06HG15G) from the

National Aeronautics and Space Administration (NASA), which facilitated the design,

manufacture, and launch of AggieSat4. These funds made possible the work of

AggieSat4, but were not directed specifically for the purpose of creating this thesis. The

contents of this document are solely the responsibility of the author and do not

necessarily represent the official views of the NASA.

vi
 NOMENCLATURE

1U Foundational CubeSat "unit", 1U = 10x10x10 cm cube

3U 3x 1U cube volume CubeSat, 3U = 10x10x30 cm block

αb Fastener coefficient of thermal expansion

αj Joint material coefficient of thermal expansion

Γ Uncertainty factor

ΔThot/cold Maximum/minimum temperature a fastener experiences

At Tensile area (surface area of threads engaged in fastener)

AGSL AggieSat Lab

AGS2 AggieSat2

AGS4 AggieSat4

B Signal bandwidth

Baud Bits per second data rate

Bevo-2 The AggieSat counterpart spacecraft from the University of Texas

ConOps Concept of operations

Cyclops Deployment mechanism for releasing payloads from ISS

D Fastener major diameter

Dpmin, ext Heli-Coil® pitch diameter

Eb Young's modulus for fastener material

Eb/No Signal to noise ratio

Ej Young's modulus for joint material

vii
EAF Experiment attachment fixture for the Cyclops deployment

mechanism

f Transmission frequency

FSU Joint ultimate shear strength

g Gravitational acceleration force (9.8 m/s2)

Grms Root-mean-square gravitational acceleration

ISIPOD 3U CubeSat canister deployment system

k Boltzmann constant (1.38E-23 J/K)

KB Fastener stiffness

Kj Joint stiffness

KTyp Nut factor

L1 Bonded material length

L2 Joint material length

Le Length of fastener thread engagement

Lhc Heli-Coil® length

Lpa Atmospheric absorption

Lpp Precipitation loss

Ls Free space loss

m Mass

Mb Megabit (106 bits)

MB Megabyte (8 Mb)

MS Margin of safety

viii
n Propagation loss

NS Number of fasteners

N0 Noise power

nL1 Joint force application distance

PAt Maximum fastener load

Pb Axial fastener load

P Fastener external load

PLDmax/min Maximum/minimum fastener preload

PSCAM The Aerospace Corporation provided camera

R Link range

RelNav Relative navigation

RELLIS Texas A&M satellite campus, formerly Riverside campus

T Torque

T0 Noise temperature

TR Receiver temperature

TP Heli-Coil® locking torque

X, Y, Z Primary Cartesian coordinate axes in given reference frame

ix
 ACRONYMS

ADCS Attitude determination and control system

AFRL Air Force Research Lab

AFSK Audio frequency shift keying

AMP Amplifier

ARD Autonomous rendezvous and docking

ARM Advanced RISC machines

BER Bit error rate

CAD Computer assisted drafting

CDH Command and data handling

CG Center of gravity

CMG Control moment gyroscope

COMM Communications

COTS Commercial off the shelf

dB Decibel

dBm Decibel milliwatt

D/L Downlink

DoD Department of Defense

DRAGON Dual RF Astrodynamic GPS Orbit Navigator

EIRP Effective isotropic radiated power

ERP Effective radiated power

x
EPS Electrical power system

FCC Federal Communications Commission

FEM Finite element method

FHSS Frequency hopping spread spectrum

FM Frequency modulation

FSK Frequency shift keying

GN&C Guidance, navigation, and control

GPS Global positioning system

G/S Ground station

HDR High data rate radio

HPB Half power beamwidth

IC Integrated circuit

IMU Inertial measurement unit

ISIS Innovative Solutions In Space

ISS International Space Station

ITU International Telecommunications Union

JAXA Japan Aerospace Exploration Agency

JEM Japanese Experiment Module

JSC NASA Johnson Space Center

LDR Low data rate

LEO Low Earth orbit

LI-ION Lithium ion (battery)

xi
LNA Low noise amplifier

LONESTAR Low Earth Orbiting Navigation Experiment for Spacecraft Testing

Autonomous Rendezvous and Docking

M2O-X Mission two objective #X

MEI Merging Excellence and Innovation Technologies

MR-X Mission requirement #X

MSC-X Minimum success criteria #X

NAS National Aerospace Standard

NASA National Aeronautics and Space Administration

NF Noise figure

NORAD North American Aerospace Defense

OA-4 Orbital ATK Cygnus Cargo Vehicle

PBER Packet bit error rate

PCB Printed circuit board

PLR Packet loss rate

PSD Power spectral density

PSI Pounds per square inch

RAM Random access memory

RCS Reaction control system

RF Radio frequency

RFIC Radio frequency integrated circuit

RMS Root mean square

xii
RPM Revolutions per minute

RW Reaction wheel

SD Secure digital memory format

SMA Subminiature version A coaxial connector

SMTRS Structures, mechanisms, thermal, and radiation systems

SNR (Eb/No) Signal to noise ratio

SOH State of health

SSIKLOPS Space Station Integrated Kinetic Launcher for Orbital Payload

Systems (now Cyclops)

STARE Space-Based Telescopes for the Actionable Refinement of

Ephemeris

STM Space Traffic Management

STS Space Transportation System (Space Shuttle)

TAMU Texas A&M University

TC Torque Coil

TNC Terminal node controller

TPI Threads per inch

TX/RX Transmit/receive

UHF Ultra-high frequency radio band

USB Universal serial bus

UT The University of Texas at Austin

UTS Ultimate tensile strength

xiii
UTJ Ultra triple junction

VDCS Visual data capture system

VHF Very high frequency radio band

VNS Visual navigation system

VSWR Voltage standing wave ratio

xiv
 TABLE OF CONTENTS

Page

ABSTRACT ...................................................................................................................... II

ACKNOWLEDGEMENTS .............................................................................................IV

CONTRIBUTORS AND FUNDING SOURCES............................................................VI

NOMENCLATURE ....................................................................................................... VII

ACRONYMS .................................................................................................................... X

TABLE OF CONTENTS ............................................................................................... XV

LIST OF FIGURES ..................................................................................................... XVII

LIST OF TABLES ...................................................................................................... XXII

1. INTRODUCTION .......................................................................................................... 1

2. DESIGN PROCESS AND RESULTING CONFIGURATION .................................... 8

2.1. Mission Requirements ............................................................................................. 8

2.1.1. NASA Defined Requirements .......................................................................... 8
2.1.2. Minimum Success Criteria ............................................................................. 13
2.2. Concept of Operations and Mission Architecture ................................................. 16
2.3. Trade Study Process .............................................................................................. 19
2.4. Spacecraft Design .................................................................................................. 21
2.4.1. Structures, Mechanisms, Thermal, and Radiation Systems (SMTRS)........... 22
2.4.2. Command and Data Handling (CDH) ............................................................ 57
2.4.3. Electrical Power System (EPS) ...................................................................... 60
2.4.4. Communications (COMM) ............................................................................ 77
2.4.5. Attitude Determination and Control System (ADCS) .................................. 104
2.4.6. DRAGON GPS (DRAGON) ........................................................................ 112
2.4.7. Visual Data Capture System (VDCS) .......................................................... 113

3. ASSEMBLY, INTEGRATION, AND TESTING ..................................................... 118

3.1. Assembly ............................................................................................................. 118

3.1.1. Prototyping ................................................................................................... 118

xv
 3.1.2. Component Assembly .................................................................................. 119
3.1.3. Soldering ...................................................................................................... 119
3.1.4. Staking and Conformal Coating ................................................................... 120
3.1.5. Locktite Adhesive ..................................................................................... 122
3.1.6. SMTRS ......................................................................................................... 123
3.1.7. CDH.............................................................................................................. 125
3.1.8. EPS ............................................................................................................... 125
3.1.9. COMM ......................................................................................................... 132
3.1.10. ADCS ......................................................................................................... 133
3.1.11. DRAGON ................................................................................................... 133
3.1.12. VDCS ......................................................................................................... 134
3.2. Integration ........................................................................................................... 134
3.3. Integrated Testing ................................................................................................ 149

4. LAUNCH, DEPLOYMENT, AND OPERATIONS.................................................. 156

4.1. Launch ................................................................................................................. 156

4.2. Deployment ......................................................................................................... 156
4.3. Operations ........................................................................................................... 165

5. CONCLUSIONS ........................................................................................................ 169

5.1. Requirements Design Evaluation ........................................................................ 169

5.2. Mission Summary ............................................................................................... 172
5.3. Spacecraft Design Summary ............................................................................... 172
5.4. Assembly Summary ............................................................................................ 173
5.5. Lessons Learned and Recommendations ............................................................ 173
5.6. Personal Contributions ........................................................................................ 176
5.7. Final Evaluation .................................................................................................. 178

REFERENCES ............................................................................................................... 180

APPENDIX A LINK BUDGET COMPONENTS AND EQUATIONS ....................... 185

APPENDIX B CUSTOM OMNETICS WIRE HARNESS EXAMPLE ....................... 196

APPENDIX C FASTENER TORQUE CALCULATION ............................................. 197

xvi
 LIST OF FIGURES

Page

Figure 1: Bevo-1 (left) and AGS2 (right)........................................................................... 4

Figure 2: Failure of Separation on Deployment (Photo by NASA) ................................... 5

Figure 3: AGS4 Deploying from ISS (Photo by NASA) ................................................... 6

Figure 4: ISIPOD CubeSat Deployment System ............................................................. 23

Figure 5: AGS4 Component Placement ........................................................................... 24

Figure 6: ISIPOD Spacer Block and I-Beams .................................................................. 25

Figure 7: Cyclops EAF and Spacer .................................................................................. 26

Figure 8: AGS4 Exterior Dimensions and Coordinate Axes ........................................... 27

Figure 9: Panel Assembly Configuration ......................................................................... 29

Figure 10: AGS4 Assembly Jig and Panel Layout ........................................................... 30

Figure 11: Solar Panel Truss Structure (inches)............................................................... 32

Figure 12: T-Shaped Solar Panel Truss (inches) .............................................................. 32

Figure 13: Long Solar Panel Truss (inches) ..................................................................... 33

Figure 14: Solar Panel Cover ........................................................................................... 34

Figure 15: Bottom Panel Mylar Covers ........................................................................ 35

Figure 16: Polycarbonate Wire Brackets ......................................................................... 36

Figure 17: Polycarbonate Wire Bracket Model ................................................................ 37

Figure 18: Clamshell Foam Lower Half........................................................................... 38

Figure 19: Clamshell Foam Upper Half ........................................................................... 39

Figure 20: Foam Packaging Housing AGS4 .................................................................... 39

Figure 21: AGS4 Launch Restraint Configuration (Image by NASA) ............................ 40

xvii
Figure 22: Top Fixed Constraints..................................................................................... 41

Figure 23: Bottom Fixed Constraints ............................................................................... 41

Figure 24: 1.5 cm Analysis Mesh..................................................................................... 44

Figure 25: 1st Mode Natural Frequency Response .......................................................... 45

Figure 26: Falcon-9 PSD Function................................................................................... 46

Figure 27: Attenuated +3 dB PSD Function .................................................................... 47

Figure 28: Solar Panel Vibration Analysis ....................................................................... 48

Figure 29: +Z Axis and Rotational Acceleration Displacement ...................................... 50

Figure 30: T-Shaped Panel Displacement ........................................................................ 50

Figure 31: Long Panel Displacement ............................................................................... 51

Figure 32: Square Panel Displacement ............................................................................ 51

Figure 33: Panel Cover Displacement ............................................................................. 52

Figure 34: AGS4 Thermal Simulation Model .................................................................. 54

Figure 35: 5 cm Mesh ....................................................................................................... 54

Figure 36: Temperature Model for β= 75° ....................................................................... 57

Figure 37: TS-7800 Flight Computer ............................................................................... 58

Figure 38: TS-SER4 Serial Expansion Board .................................................................. 59

Figure 39: CDH Stack ...................................................................................................... 59

Figure 40: AGS4 Power System Diagram ....................................................................... 61

Figure 41: EPS and Battery Installed on AGS4 ............................................................... 61

Figure 42: Battery Charge Regulator Board .................................................................... 62

Figure 43: Voltage Regulator Board ................................................................................ 63

Figure 44: Torque Coil Control Board Layout ................................................................. 63

Figure 45: Output Board .................................................................................................. 64

xviii
Figure 46: Hazard Control Board ..................................................................................... 65

Figure 47: EPS Controller Board ..................................................................................... 66

Figure 48: Battery Prototype ............................................................................................ 67

Figure 49: Battery Cut-Away View ................................................................................. 68

Figure 50: Battery Box Vent and Cover........................................................................... 69

Figure 51: Battery Protection Circuit ............................................................................... 70

Figure 52: Assembled Battery Box .................................................................................. 70

Figure 53: Top and Side Solar Panel Configuration ........................................................ 71

Figure 54: Bottom Face Solar Panel Configuration ......................................................... 72

Figure 55: Inhibit "Active" ............................................................................................... 74

Figure 56: AGS4 Inhibit Switches ................................................................................... 75

Figure 57: Prior AGS4 Foot Switch Design..................................................................... 76

Figure 58: Previous Design Switch Preparation .............................................................. 76

Figure 59: Kenwood TH-D72A ....................................................................................... 87

Figure 60: HDR Radio Unit Design (photo by Graves, Dec. 2011) ................................ 90

Figure 61: SM04093-36HS Amplifier ............................................................................. 91

Figure 62: Digi XTend .................................................................................................. 93

Figure 63: Crosslink Antenna Pattern .............................................................................. 95

Figure 64: HDR Antenna Pattern ..................................................................................... 96

Figure 65: Antennas Installed on AGS4 (Photo by NASA)............................................. 97

Figure 66: Sun Sensor Installed ..................................................................................... 106

Figure 67: Torque Coil ................................................................................................... 110

Figure 68: PSCAM ......................................................................................................... 114

Figure 69: Camera Location (red) .................................................................................. 115

xix
Figure 70: PSCAM Full Resolution ............................................................................... 116

Figure 71: PSCAM Video Resolution ............................................................................ 117

Figure 72: PSCAM Thumbnail ...................................................................................... 117

Figure 73: Vacuum Jar Setup ......................................................................................... 121

Figure 74: Conformal Coat Visible on Battery Protection Circuit ................................. 122

Figure 75: Solar Panel Cover ......................................................................................... 124

Figure 76: Double Sided Kapton Cut-Out .................................................................. 126

Figure 77: Solar Panel Jigs ............................................................................................. 127

Figure 78: Solar Cells in Assembly Jig .......................................................................... 128

Figure 79: Assembled Solar Panel ................................................................................. 129

Figure 80: Battery Assembly.......................................................................................... 130

Figure 81: Prototype and Flight Battery Boxes .............................................................. 131

Figure 82: -Z Panel Assembly Jig .................................................................................. 135

Figure 83: Cyclops EAF and Inhibit Switches ............................................................... 136

Figure 84: CDH Integration ........................................................................................... 137

Figure 85: AGS4 Pre Fold Up ........................................................................................ 138

Figure 86: Illustration of +Y Panel Fold Up .................................................................. 139

Figure 87: Folded Up Side Panels with Handles on Assembly Jig ................................ 140

Figure 88: +Z Assembly Jig Panel ................................................................................. 141

Figure 89: Illustration of +Z Jig Attachment ................................................................. 141

Figure 90: Illustration of +Z Panel Fold Up ................................................................... 143

Figure 91: Bus with Solar Panels and Covers Installed ................................................. 144

Figure 92: Foam Fit Check............................................................................................. 145

Figure 93: -Z Solar Panels Installed ............................................................................... 146

xx
Figure 94: AGS4 Test Stand (behind AGS4) ................................................................. 147

Figure 95: ISIPOD Payload Access Connectors ............................................................ 151

Figure 96: Computer Access Ports for AGS4 (Circular) and Bevo-2 (DB-9/DB-15) ... 152

Figure 97: Human Factors Testing ................................................................................. 153

Figure 98: Dynamic Random Vibration Testing ............................................................ 154

Figure 99: Astronauts Removing Lower Solar Panel Covers (Photo by NASA) .......... 157

Figure 100: AGS4 Removed from Packaging (Photo by NASA) .................................. 157

Figure 101: AGS4 Install onto Cyclops Table (Photo by NASA) ................................. 158

Figure 102: AGS4 Solar Panel Covers Removed and Antennas Installed (Photo by
NASA) ............................................................................................................ 159

Figure 103: Inhibit "Active" (Photo by NASA) ............................................................. 160

Figure 104: AGS4 Insertion into JEM Airlock (Photo by NASA) ................................ 160

Figure 105: JEM Airlock (Photo by NASA) .................................................................. 161

Figure 106: AGS4 and Cyclops Table Moved onto Porch (Photo by NASA) ............... 162

Figure 107: Dextre Moving AGS4 to Deploy (Photo by NASA) .................................. 162

Figure 108: AGS4 Pre-Deploy (Photo by NASA) ......................................................... 163

Figure 109: AGS4 Release (Photo by NASA) ............................................................... 164

Figure 110: AGS4 Free Flying (Photo by NASA) ......................................................... 165

Figure 111: Signatures and Aggie Rings of AGSL Members and NASA Management
displayed on AGS4. Also shown (top) is the Aggie Ring of Civil
Engineering graduate Patrick Brand (’81) whose father, Vance D. Brand,
carried it into space as commander of Space Shuttle Columbia Mission
STS-5 in November 1982. .............................................................................. 179

Figure 112: Battery Box Thermal Analysis Probing ...................................................... 200

Figure 113: Fastener Dimension Diagram ..................................................................... 201

xxi
 LIST OF TABLES

Page

Table 1: AGS4 Mission Objectives .................................................................................... 9

Table 2: Derived Mission Requirements .......................................................................... 10

Table 2: Derived Mission Requirements (Continued) ..................................................... 11

Table 2: Derived Mission Requirements (Continued) ..................................................... 12

Table 2: Derived Mission Requirements (Continued) ..................................................... 13

Table 3: Minimum Success Criteria ................................................................................. 14

Table 3: Minimum Success Criteria (Continued) ............................................................ 15

Table 3: Minimum Success Criteria (Continued) ............................................................ 16

Table 4: Comm Frequency Band Trade Study Example .................................................. 20

Table 5: Mass Configurations .......................................................................................... 28

Table 6: Center of Gravity Location ................................................................................ 28

Table 7: Frequency Modes for Mesh Sizes ...................................................................... 43

Table 8: Falcon-9 Random Vibration Environment ......................................................... 45

Table 9 : Solar Panel Vibration Factor of Safety ............................................................. 48

Table 10: Axial and Rotational Acceleration Results ...................................................... 49

Table 11: Panel Acceleration Stress and Displacement ................................................... 52

Table 12: AGS4 ISS View Factors .................................................................................. 55

Table 13: Radiative Heat Flux Sources and Sinks ........................................................... 55

Table 14: AGS4 Power Budget ........................................................................................ 60

Table 15: AGS4 State of Health Packet ........................................................................... 80

Table 16: Health Data Volume......................................................................................... 81

xxii
Table 17: DRAGON GPS Data Generated ...................................................................... 81

Table 18: Visual Data Downlink Volume ........................................................................ 82

Table 19: ADCS Data Packet Size ................................................................................... 83

Table 20: Total ADCS Data Generation .......................................................................... 83

Table 21: Total Downlink Data ........................................................................................ 84

Table 22: HDR and LDR Downlink Capacity ................................................................. 85

Table 23: AGS4 Mission Downlink Times ...................................................................... 86

Table 24: AGS4 Beacon Format ...................................................................................... 89

Table 25: Crosslink Data Packet ...................................................................................... 93

Table 26: Uplink Link Budget ......................................................................................... 99

Table 27: LDR Downlink Link Budget .......................................................................... 100

Table 28: HDR Downlink Link Budget ......................................................................... 101

Table 29: Crosslink Link Budget ................................................................................... 102

Table 30: Torque Coil Construction............................................................................... 111

Table 31: PSCAM Imaging Capabilities ........................................................................ 114

Table 32: LDR VHF Noise Temperature ....................................................................... 192

Table 33: LDR UHF Noise Temperature ....................................................................... 193

Table 34: HDR Noise Temperature................................................................................ 194

Table 35: Crosslink Noise Temperature ......................................................................... 195

Table 36: Fastener Properties ......................................................................................... 197

Table 37: Heli-Coil Properties .................................................................................... 197

Table 38: Joint Properties ............................................................................................... 198

Table 39: Battery Box Thermal Preload ........................................................................ 202

Table 40: Uncertainty Factor Γ ...................................................................................... 203

xxiii
Table 41: Joint Calculation Results ................................................................................ 205

Table 42: Preload and Margin of Safety for Fastener Torque ........................................ 206

xxiv
 1. INTRODUCTION

Satellites play a crucial role in modern society, facilitating global

communications, entertainment, navigation, security services, and countless other

capabilities that most people take for granted on a day-to-day basis. As resources for

spacecraft are always constrained, the global trend is toward designing and

manufacturing spacecraft that are smaller, while being as capable if not more so than

their predecessors. To give some perspective, more traditional spacecraft have been on

the order of 1000-5000 kg with dimensions in the 10s of meters, and current thinking is

toward spacecraft on the order of 300 kg or less with body dimensions under 2 meters.

Leadership at the Defense Innovation Unit and Air Force Research Lab (AFRL) believe

that the current market trend is to utilize large constellations of small satellites to provide

both government and commercial services [1]. A distributed system of small spacecraft

may offer new mission concepts to meet requirements, may offer advantages in

economies of scale related to manufacturing and the supply chain, may offer advantages

in launch costs and available opportunities, is less prone to single point of failure issues,

can be incrementally improved as technology advances, can be replenished in the event

of single-unit failures, and is more hardened against adversary actions than a single

highly capable relatively large spacecraft.

With recent advances by nations such as China and Russia, the unparalleled

advantage that the United States has held in space is being steadily eroded [2].

Moreover, as of today, some 72 nations have a space program, and eight countries

1
along with the European Space Agency have the ability to launch objects into space [3].

In response, the Department of Defense has adopted a space policy to leverage the

unmatched innovation of the American commercial space industry to provide solutions

as space becomes more and more congested, contested, and competitive [2, 4, 5]. This

strategy requires that the United States engages the best talent. Satellite design labs at

universities provide some of the most effective means of training students to be

competent engineers, ready to enter the workforce and contribute to the solutions of

utilizing space to improve the lives of people around the world, and to challenge the

actions of bad actors.

Universities have been leaders in developing small satellites and small satellite

technologies. Prior to 2015, the roughly 75% of microsatellites (spacecraft under 100

kg) and nanosatellites (spacecraft under 10 kg) launched and operated were from

University programs, and it was not until 2015 that commercial industry began to field

the majority of small vehicles [6]. The CubeSat standard developed by California

Polytechnic State University (Cal-Poly) San Luis Obispo had a dramatic effect towards

standardization in this industry, defining spacecraft by size and mass constraints [7].

This standardization has led to a significant increase in the number of small satellites

launched each year [6]. Very common sizes have been "1U" (10x10x10 cm) and "3U"

(10x10x30 cm), though larger sizes exist [8].

Texas A&M University's AggieSat Lab (AGSL) has worked with NASA, the

AFRL University Nanosatellite competition, and Lawrence Livermore National

Labs/Naval Postgraduate School in various projects in small spacecraft design. Under

2
the NASA program, AGSL produced two small spacecraft to test navigation experiments

and train University students on spacecraft design and fabrication [9]. With Lawrence

Livermore National Labs, AGSL participated in the design and engineering analysis of

the STARE (Space-based Telescopes for the Actionable Refinement of Ephemeris) 3U

CubeSats in , an orbital debris tracking demonstration and space traffic management

(STM) experiment [10][12][13]. AGSL continues to participate in the AFRL University

Nanosatellite competition, with its most recent entry being AggieSat 6 whose planned

missions are the demonstration of an on-orbit satellite locator and the collection of

radiation data.

AGSL is a student led spacecraft design laboratory, with a focus on developing

the skill sets of students. In this endeavor, AGSL has had the opportunity to design,

build, launch, and operate two small spacecraft through a partnership with the National

Aeronautics and Space Administration (NASA) through the LONESTAR program.

LONESTAR (Low Earth Orbit Navigation Experiment for Spacecraft Testing

Autonomous Rendezvous and docking) was a NASA program to partner the student

spacecraft labs of TAMU and the University of Texas at Austin (UT) to produce a series

of spacecraft that incrementally developed skills, methods, and techniques to culminate

in Autonomous Rendezvous and Docking (ARD) between spacecraft from both labs.

Both Universities each flew two spacecraft missions as part of the program.

The first pair of spacecraft were AggieSat2 (AGS2) and Bevo-1, both 5 in. DoD

standard CubeSats deployed from Space Shuttle Endeavor on STS-127 (deployed July

15, 2009, re-entry March 17, 2010). The purpose of these two spacecraft was to give

3
each lab experience in building spaceflight hardware, as well as to test a new GPS

receiver developed by NASA. Due to a possible stabilizing magnet polarity

misalignment and a spring loaded antenna, the two spacecraft did not completely

separate after deployment from their spring-loaded canister. While the two did not

separate completely, the spring separation mechanism did force them apart sufficiently

to remove the AGS2 electrical inhibit and allow for it to activate. Unfortunately, Bevo-1

never turned on. Even though the antenna was unable to fully deploy, reduced ground

communication was still possible, and health status and GPS data were able to be

recorded and transmitted for much of the flight. The two spacecraft remained conjoined

until their re-entry according to NORAD.

Figure 1: Bevo-1 (left) and AGS2 (right)

4
 Figure 2: Failure of Separation on Deployment (Photo by NASA)

The second mission consisted of spacecraft named AggieSat4 (AGS4) and Bevo-

2. The purpose of this mission was to again test the GPS receiver model, as well as to

perform relative navigation and attitude control demonstrations. AGS4 was a larger 50

kg class vehicle, while Bevo-2 was a standard 3U CubeSat to be deployed from AGS4.

Both were launched to the International Space Station (ISS) as a combined unit aboard

Orbital OA-4, and deployed from the Station on January 29, 2016.

5
 Figure 3: AGS4 Deploying from ISS (Photo by NASA)

Unfortunately, technical problems prevented the mission demonstration from

occurring. Bevo-2 was released without being commanded, and two separate objects

were tracked by NORAD. As the release system was three fault tolerant and had been

tested thoroughly, the suspected cause of this release was mechanism failure with the

commercial off the shelf (COTS) payload deployer due to repeated thermal cycles

beyond what the deployer had been tested to by the manufacturer. AGS4 also

experienced problems in that the computer memory would fill up after ~52 hours,

6
forcing a reset. Beacons stopped being received after about 7 days after deployment,

and this situation was unresolved at the time of re-entry on March 12, 2018.

The subsequent sections of this document detail the design of the mission and

spacecraft, assembly, testing, and operation of AGS4. Additionally, many lessons

learned and other information are provided that will be helpful to AGSL students as they

prepare to design, build, and fly spacecraft of their own.

7
 2. DESIGN PROCESS AND RESULTING CONFIGURATION

2.1. Mission Requirements

Mission requirements are the driving factor shaping the design of any satellite or

its mission. They are the means by which the customer (NASA) defines what they want

to happen. These definitive requirements are then refined by both the design team and

the customer into a set of derived requirements that dictate how the mission is designed,

what the spacecraft will do, and how it will be designed to do these things.

One note regarding the verbiage of the requirements is that they use the word

"will", where they should instead use "shall". While many may view the words as

interchangeable, from personal experience working with the United States government

since the end of this project, it has been learned that "shall" is the preferred word as

"will" implies some degree of uncertainty to some groups. For the purposes of this

thesis, the language in the requirements was left as agreed to at the time by all parties for

this project in order to document the requirements as worked to.

2.1.1. NASA Defined Requirements

The requirements provided by NASA were in the form of mission objectives, and

are outlined in the table below. These mission objectives applied to both the AGSL and

UT teams, as the two were partners for these missions. As the AGS4 mission was the

second in the LONESTAR campaign, they were given an M2O designation.

8
 Table 1: AGS4 Mission Objectives
- Mission 2 Objectives (M2O)
M2O-1 Both teams will evaluate sensors including but not limited to: Global
Positioning System (GPS) receivers, Inertial Measurement Units (IMU),
rate gyros, and accelerometers.
M2O-2 Both teams will evaluate a 1st generation Reaction Control System (RCS).
M2O-3 Both teams will work together to evaluate a 1st generation Guidance,
Navigation and Control (GN&C) system including guidance algorithms,
absolute navigation, and relative navigation.
M2O-4 Both teams will work together to evaluate communications capabilities
between two spacecraft and from each spacecraft to their ground stations.
M2O-5 Both teams will evaluate the capability to take video.
M2O-6 Each University will select two members to create and organize an
interface monitoring function in order to maintain robust and productive
communications between the university teams and oversight members:
NASA and MEI Technologies.
M2O-7 Each spacecraft will have three-axis stabilization.
M2O-8 Both teams will work together to evaluate GN&C system interfaces and
compatibility and the testing required for these systems.

Using these mission objectives, the two teams and NASA management derived

the following set of mission requirements which were used to shape the design of AGS4

and the mission.

Mission requirements often trace back to the set of customer defined

requirements (M2O's) to show that they are being met. In this case, each M2O should

have at least one MR tracing to it to show that the requirement is being met. However,

some other derived requirements are more operational constraints such as the envelope

and mass constraints outlined below, which were dictated by the available NASA

resources provided to the program. The table below details the derived mission

requirements, and how they trace to the M2O.

9
 Table 2: Derived Mission Requirements
Mission Requirements Trace- AGSL Verification
ability Process
- Satellite Envelope Requirements
MR-1 The entire LONESTAR envelope is Analysis- evaluate
not to exceed the maximum CAD model
dimensions of the NASA deployment Lab testing- measure
system for the Japanese Aerospace assembled structure
Exploration Agency (JAXA)
Japanese Experiment Module (JEM)
airlock on the International Space
Station (ISS).
- Satellite Mass
MR-2 The entire LONESTAR mass is not Lab testing- weigh
to exceed 100 kg. integrated structures
MR-3 AggieSat4 shall not exceed 50 kg. Analysis- evaluate
CAD model
Lab testing- weigh
assembled structure
MR-4 Bevo-2 shall not exceed 35 kg. -
MR-5 This project will follow the standard Documentation-
project management schedule. Integrated project
schedule
MR-6 Each University team will select two M2O-6 Enforced by lab
members to create and organize an management.
interface monitoring function in order
to maintain robust and productive Email communication
communications between the and weekly telecoms
university teams and oversight between both
members: NASA and MEI Universities and JSC.
Technologies.

10
 Table 2: Derived Mission Requirements (Continued)
Mission Requirements Trace- AGSL Verification
ability Process
- GN&C System
MR-7 Each spacecraft will have three-axis M2O-1 Analysis-
stabilization capability and M2O-2 simulate space
demonstrate it. M2O-7 environment and
control disturbances in
Simulink
On orbit-
attitude sensor data will
be gathered while
AGS4 tracks Bevo-2.
MR-8 Both teams will work together to M2O-1 Lab testing-
evaluate any GN&C system M2O-2 Crosslink testing
interfaces and compatibility and the M2O-3 exchanges navigation
testing required for these systems. M2O-8 solutions.
MR-9 Each team will evaluate and test an M4O-1 Analysis-
additional component that will be Ensure that design
needed for the next generation of includes components to
their spacecraft and the final mission. test next generation
capability.
MR-10 Each team will evaluate and test M2O-2 On orbit-
portions of their future generation M2O-3 AGS4 will demonstrate
control system algorithms. M3O-2 active tracking
M3O-3 capability.
- Communications
MR-11 Each spacecraft will communicate M2O-4 Lab testing-
with the ground station, both simulate mission
directions. communications using
ground station
hardware
On orbit-
AGS4 will send packets
and receive commands
from the Ground
Support team

11
 Table 2: Derived Mission Requirements (Continued)
Mission Requirements Trace- AGSL Verification
ability Process
MR-12 Each spacecraft will communicate M2O-4 Lab testing-
with the other when separated. M2O-8 Joint simulated
crosslink testing will be
done prior to
integration
On orbit-
AGS4 will crosslink
navigation information
with Bevo-2 after the
release event
MR-13 Each satellite will exchange its GPS M2O-1 Lab testing-
solution with the other satellite. M2O-4 Joint simulated
M2O-8 crosslink testing with
GPS packets will be
done prior to
integration
On orbit-
AGS4 will crosslink its
GPS solution
information with Bevo-
2 after the release event
MR-14 Each spacecraft will evaluate the M2O-5 Lab testing-
viability and capability of AGS4’s camera will
downloading captured visual capture an image with
evidence. its camera and
downlink the image to
the ground station
hardware
On orbit-
AGS4 will capture
visual evidence of the
release event and
downlink the images to
the Ground Support
team
12
 Table 2: Derived Mission Requirements (Continued)
Mission Requirements Trace- AGSL Verification
ability Process
- Payloads
MR-15 Each spacecraft will carry a GPS M2O-1 Lab testing-
system for use in determining the M2O-2 a GPS system will be
viability of navigation solutions for M2O-3 used
the final mission. M2O-8
M4O-1
MR-16 Each spacecraft will downlink GPS M2O-1
data for the NASA team to be able to M2O-2
evaluate the operation of the receiver. M2O-3
M2O-4
- General Capabilities
MR-17 The two spacecraft will separate from
each other.
MR-18 Each spacecraft will take and M2O-4
downlink video or still photographs. M2O-5
MR-19 Once separated, each spacecraft will M2O-1
provide Relative Navigation M2O-2
(RelNav) solutions. M2O-3
M2O-4
M2O-8
M3O-2

2.1.2. Minimum Success Criteria

With these derived mission requirements agreed to by both satellite teams and

NASA management, the two teams then further refined the MR into a series of

Minimum Success Criteria (MSC). These are codifications of what each spacecraft will

do in order to meet the MR, and define the mission concept of operations. These derived

MSC's are outlined in the table below.

13
 Table 3: Minimum Success Criteria
- Minimum Success Criteria (MSC) Trace-
ability
MSC-1 The LONESTAR-2 spacecraft will demonstrate the on-orbit MR-7
performance of the three-axis stabilization systems to an
accuracy described by MSC 1.1 and MSC-1.2 for the Bevo-2
and AggieSat4 the spacecraft, respectively.
MSC- Bevo-2 will conduct two different confirmations: image and 3- MSC-1
1.1 axis stabilization confirmation. Image confirmation will
consist of pointing at an object and capturing an image 150
seconds apart, then downlinking both images. 3-axis
stabilization will consist of recording all ADCS data and
downlinking all pertinent information.
MSC- The AggieSat4 spacecraft will demonstrate the on-orbit MSC-1
1.2 performance of its three-axis stabilization system to an
accuracy adequate to accomplish one of the maneuvers
contained within MSC-3.2.
MSC-2 Each LONESTAR-2 spacecraft will demonstrate the on-orbit MR-9
performance of its “additional component”.
MSC- Bevo-2 shall verify basic functionality and characterize the MSC-2
2.1 performance of the cold-gas thruster system using GPS and
IMU measurements.
MSC- AggieSat4 will downlink carrier phase GPS data from two MSC-2
2.2 onboard GPS antennas. This will enable an accurate RelNav
vector between the two antennas to be computed and analyzed
on the ground.
-OR-
AggieSat4 will demonstrate enhanced communications speeds
for the downlinking of video and image data.
MSC-3 Each LONESTAR-2 spacecraft will demonstrate the on-orbit MR-10
performance of the future generation control system
algorithms.

14
 Table 3: Minimum Success Criteria (Continued)
- Minimum Success Criteria (MSC) Trace-
ability
MSC- Bevo-2 shall demonstrate minimum success of the future MSC-3
3.1 generation control system algorithm by performing one of the
following two actions:
1) After initialization, Bevo-2 shall actively point the
Visual Navigation System (VNS) towards the
AggieSat4 using crosslinked GPS relative position
solutions with the performance data downlinked for
evaluation.
2) A series of preprogrammed reorientation and
translational maneuvers shall be executed with the
performance data downlinked for evaluation.
MSC- AggieSat4 will demonstrate future generation control system MSC-3
3.2 algorithms with the successful execution of one of the
following maneuvers, with the first being the primary
objective:
1) AggieSat4 will actively track Bevo-2 using crosslinked GPS
data until Bevo-2 is out of range for crosslink.
2) AggieSat4 will actively track dummy propagated GPS data
once every second for a period of time equivalent to Bevo-2
being “out-of-range” based on the release speed and direction.
3) AggieSat4 will hold a target object (such as the Moon and
Earth) in the field of view of the spacecraft’s camera for a
minimum of fifteen (15) seconds. Photographs of the object
with timestamps should be downloaded to confirm.
MSC-4 Each LONESTAR-2 spacecraft will be capable of receiving MR-11
commands, acting upon those commands, and downlinking a
response within a reasonable amount of time.
MSC-5 Each LONESTAR-2 spacecraft will send to and receive from MR-8
the other, and then validate a “dummy” DRAGON GPS MR-12
preloaded data file after separation.
MSC-6 Each LONESTAR-2 spacecraft will send to and receive from MR-8
the other, and then assess fifty (50) GPS state solutions. MR-12
MR-13
MR-15
MR-19
15
 Table 3: Minimum Success Criteria (Continued)
- Minimum Success Criteria (MSC) Trace-
ability
MSC-7 Each LONESTAR-2 spacecraft will collect and downlink MR-14
images at a minimum resolution of 1024x768. MR-18
MSC-8 Each LONESTAR-2 spacecraft will collect and downlink at MR-15
least two (2), not necessarily consecutive, orbits’ worth of MR-16
DRAGON GPS data for delivery to the NASA DRAGON GPS
development team.
MSC-9 Each spacecraft will confirm the separation event between the MR-8
two LONESTAR-2 spacecraft through direct observation. MR-14
MR-17
MR-18
MSC-10 Each LONESTAR-2 spacecraft will exchange 50 RelNav MR-8
solutions with the other. MR-12
MR-13
MR-15
MR-19
MSC-11 Each LONESTAR-2 spacecraft will downlink complete state MR-11
of health data to the ground at a rate which constitutes a MR-20
reasonable amount of time between transmissions.

2.2. Concept of Operations and Mission Architecture

Using the MR and MSC as the shaping forces, the AGSL and UT teams

developed a concept of operations (ConOps) for the mission, outlining what tasks should

be performed and in what order. The mission profile begins with AGS4 being launched

to the ISS as soft-stowed cargo. After arriving at the ISS, AGS4 would then be

unpacked by the astronauts. Using the Cyclops (formerly SSIKLOPS, Space Station

Integrated Kinetic Launcher for Orbital Payload Systems) Experiment Attachment

Fixture (EAF), AGS4 would be locked onto the Cyclops deployment table, which would

release the payload from the end of a robot manipulator arm. Cyclops would exit the
16
ISS via Japanese JEM airlock, where the robot arm would maneuver it into place before

activating the release mechanism.

Upon release, solar power would activate the power system, which would begin

the boot-up of the computer system. Additionally, EPS would contain a module that

houses three separate timer circuits to prevent Bevo-2 from being released while too

close to ISS. All three timers would have to reach or exceed the value in order for the

computer to be able to command the Bevo-2 release.

After initial boot up, AGS4 would collect and process telemetry and health status

information, and transmit it in a beacon once every six seconds. This data would

provide a snapshot of the current status of AGS4 components, as well as provide a

location indicator for tracking purposes. Using these beacons, communications would

be established by the ground station, and all state of health data downloaded. AGS4

would then be commanded to begin its collection of the two orbits of GPS data per

MSC-8 for download at the next ground pass.

After downlinking the GPS data, which would take multiple ground passes to

download, both the AGSL and UT teams would prepare to initiate the RelNav

experiment. AGS4 would then be commanded to begin the RelNav process. After

ensuring that Bevo-2's batteries were topped off, AGS4 would orient to 90° off of the

velocity vector and release Bevo-2, as this release vector provides the most time in radio

range with each other and several re-contact opportunities. AGS4 would use a camera to

image the separation. The operations to this point would take sufficient time that the

timers would no longer inhibit the deployment.

17
 After deployment, the inhibit switches on Bevo-2 would close and allow it to

boot-up. Crosslink radios would be used to exchange GPS solutions between the

spacecraft. As GPS receivers can take several minutes to converge on a navigation

solution, especially at orbital speeds, the crosslinked AGS4 solution would be used to

"hot start" the Bevo-2 receiver. Once both spacecraft achieve navigation solutions,

AGS4 and Bevo-2 would use the location of the other to calculate the relative position of

the other with respect to themselves, and point and track the other as they progress

through the orbit. This scenario would demonstrate both RelNav and attitude control.

In case something went wrong with the crosslinking, AGSL developed a

contingency plan to track a simulated Bevo-2 using a propagated solution, and exercise

point-and-hold maneuvers. AGS4 would also image the Earth and Moon for a visual

confirmation of pointing stability.

Primary or contingency location data for both spacecraft, and attitude data and

imagery from AGS4 would be downloaded over the next several orbits, as the expected

data generated would exceed what can be expected in a single pass. The position and

pointing data would be used to verify that AGS4 tracked Bevo-2, and the images would

verify the separation.

After all RelNav experiment data was downlinked, the remaining life of the

spacecraft would be spent downlinking extra captured images, the complete state of

health record, and practicing further control maneuvers by imaging Earth and other

celestial objects.

18
2.3. Trade Study Process

A trade study is a means of comparing options for processes, systems, or

individual components against others of their type in order to determine a solution that

best fulfills the mission requirements. Sometimes the most capability is not best for the

job at hand, and "better is the enemy of good".

For all things being compared, the trade study begins by listing all options being

considered to fulfill the mission requirements, along with their attributes being

compared. A table format can help for visualization.

First, any options that do not meet any of the mission requirements are removed.

Then, the attributes of each option are compared to those of the other candidates, and

ranked. The designers will need to use the mission requirements as a guide to determine

which attributes carry the most weight. Color coding cells can be used if helpful. The

option that has the most positives and fewest drawbacks is usually apparent after

ranking. However, some subjective factors such as personnel experience with a

particular option, or company reputation and past experience should be considered too as

they can produce a smoother path to the end goal, which can definitely be worth not

going with the "perfect" approach on paper.

The example covered below outlines the process followed to determine in which

radio band AGS4 should seek a license to operate. Subsequent trade studies were then

performed in the same manner to determine the hardware selection. This same process

is equally applicable to all subsystems and other areas of the spacecraft design decision-

making process.

19
 The Communications team considered three different frequency bands to use for

AGS4. These were Commercial UHF/S-band, Experimental S-band, and Experimental

Amateur VHF/UHF. The frequency bands had differing characteristics as laid out in the

table below.

Table 4: Comm Frequency Band Trade Study Example

Commercial Experimental S- Experimental
UHF/S-band band Amateur
VHF/UHF
License Cost $50,000+ $65 $65
License Lead Time 6-12 mo. 3-6 mo. 3-6 mo.
Equipment Cost $100's-$1000's $100's-$1000's $100's
Worldwide Pay-per-use Limited Common, Amateur
Receiver radio operator
Availability volunteers
Dedicated Yes No, shared with No, shared with
Frequency? other users other users
AGSL Ground Yes Yes Minimal
Station
Modifications
required?
Encryption Legal Yes No No
Signal Propagation Low Low Very Low
Loss
Interference Level Very Low Very Low Very Low
at Ground Station

Commercial frequencies, while benefiting from not requiring the sharing of the

airwaves with other users, were almost immediately discarded due to the high regulatory

cost, which would have placed significant pressure on the program budget. Choosing

between the remaining two was not difficult, as the large number of amateur operators

worldwide willing to provide communications services, and not needing to modify the

AGSL ground station were huge positive factors. The largest drawback to using

20
frequencies that overlap with Amateur bands is the potential for interference at the

receiving station. Experience with AGS2, and measurements in the rural setting of the

ground station on the Riverside Campus (now RELLIS) indicated that the risk of

interference was unlikely and could be worked around.

In this manner, the AGS4 team determined that using Amateur VHF (144 MHz)

and UHF (440 MHz) bands with Experimental authorization would produce a

communications system with the greatest chance of successful throughput and mission

data retrieval.

2.4. Spacecraft Design

Engineering design is a multi-iteration process to determine a vehicle platform

design that meets all MR and MSC. The design is often split into subsystem engineering

teams responsible for an aspect of the vehicle. The process is iterative because no

subsystem is an island, and any change to one of them can have significant repercussions

on all the rest. For instance, if it is determined that a navigation sensor or antenna is

located in a position where it will be ineffectual, moving it to a better location could

impact the number of solar panels that can be fitted to that face, and as a result every

system now has less power with which to accomplish its tasks. Almost every change or

adjustment has downstream effects, particularly to the vehicle mass and center of gravity

(CG). The following sections outline the six spacecraft subsystems and how they were

developed into the final flight configuration.

The decision was made to use U.S. Customary units for the design of mechanical

systems and hardware on AGS4, and to use the Metric system for analysis and

21
calculations. This decision was made because the majority of machine shops and tools

available in the United States use Customary units, making it far easier for AGSL to

source parts, have pieces machined, and source fasteners. Metric units were used on

analyses as students were more familiar with them for engineering calculations from

coursework. This arrangement was acceptable to the NASA management as it is

commonly implemented at NASA, and was reflected in the documentation and data

shared with AGSL. The following sections will present quantities such as mechanical

dimensions and fastener torque using Customary units, and results from analyses such as

displacements, temperatures, and mesh sizes are in Metric (though sometimes both are

indicated for clarity).

2.4.1. Structures, Mechanisms, Thermal, and Radiation Systems (SMTRS)

The main role of the SMTRS team was to design and analyze the AGS4 structure

to include bus design and component locations, as well as thermal and loads analysis.

2.4.1.1. Bus Design

AGS4 was designed to host the COTS ISIPOD CubeSat dispenser provided by

UT to release Bevo-2 for the RelNav experiment. This design constraint was central to

the design process of AGS4, as the bus was practically built around the ISIPOD.

22
 Figure 4: ISIPOD CubeSat Deployment System

Since the beginning of the project, the AGS4 design has been a box shape of the

approximate maximum dimensions allowable by the Japanese JEM airlock. There have

been several design iterations as subsystem designs mature, all focusing on maintaining

the necessary balance for the attitude control system to function properly. This was

accomplished by a combination of logic, trial, and error, arranging the subsystem

components and wiring in different configurations until the center of gravity (CG) was in

the appropriate zone. The Cyclops deployment mechanism required that the AGS4 CG

fall within 0.25" from the center of the EAF in the X and Y axes, with no restriction on

the Z axis. The figure below outlines the component placement from a top view that

achieved this balance.

23
 Figure 5: AGS4 Component Placement

It was also important to have the CG close to the center of the force vector

created by Bevo-2's release, as being significantly off would impart a torque to AGS4

that could rotate the vehicle such that the camera no longer had Bevo-2 in the field of

view. ADCS determined that the AGS4 CG would need to be within 0.5" from the

ISIPOD force vector to tolerate the rotational forces. Furthermore, it was desirable to

have the spacecraft CG be as close to the geometric center as possible so that the vehicle

dynamics changed the least amount after Bevo-2 deployment. In order to achieve this

CG location, it was necessary to raise up the ISIPOD launcher such that it met the CG,

as lowering the CG was not practical past a certain point. In order to preserve the mass

budget and control authority, this elevation block would need to be lightweight. With

24
significant cut outs, the spacer block was light enough and strong enough to bring Bevo-

2 up to the CG line. The spacer block is shown in the figure below.

Figure 6: ISIPOD Spacer Block and I-Beams

25
 Figure 7: Cyclops EAF and Spacer

Also seen in Figure 6 above are two "I-beam" structural supports. These

supports were added to strengthen the vehicle as initial analysis indicated that the +Z

face would sag significantly under launch loads, and the natural frequency was closer to

that of the launch vehicle.

The final result of AGS4 was a vehicle with a deployment envelope of 32.2" x

29.3" x 20.2" including all protrusions and attachments such as antennas and the Cyclops

EAF. The positions of these protrusions were such that they still satisfied the airlock

envelope clearance requirements [14]. The motion of the airlock door is such that a box

of these dimensions would not fit, but limited protrusions on certain sides do not cause

interference. The figures below outline the bus and protrusion dimensions, as well as the

26
coordinate system directions. In this coordinate system, the panel with the EAF is the -Z

face, and the panel with the ISIPOD protruding out is the -X face.

Figure 8: AGS4 Exterior Dimensions and Coordinate Axes

The net result of all this component placement and structural components was a

vehicle mass and CG properties as outlined in the tables below. The AGS4 portion of

the launch mass was 48.8 kg, satisfying the requirement to remain under 50 kg. This

was calculated by creating a mass budget, which is a list of all structures, components,

wires, fasteners, and estimates for epoxies and coatings by their quantities and masses to

ensure that the project stays under the required amount.

27
 Table 5: Mass Configurations
Mass Configurations
Launch Configuration 52.25 kg 116.3 lb
Cyclops Deployment 51.62 kg 113.8 lb
Configuration
With Solar Panel Covers, 48.8 kg 107.6 lb
without Bevo-2
Without Solar Panel Covers, 47.67 kg 105.1 lb
without Bevo-2

Four different mass configurations were evaluated to ensure that the

requirements were being met, as well as to determine the dynamic properties for active

spacecraft pointing control. The first was the launch configuration, which verified that

the combined spacecraft mass was under 100 kg. The second was the mass of AGS4 and

Bevo-2 without the solar panel covers, as it would be released from the ISS and begin

early operations. The third was AGS4 with the solar panel covers to verify that all

AGSL hardware going to space was under the 50 kg limit. And fourth was the

configuration for AGS4 free flying after Bevo-2 release.

Table 6: Center of Gravity Location

Center of Gravity (Cyclops EAF
Coordinates)
X -0.082"
Y 0.142"
Z 7.588"

The final component configuration placed the CG within the 0.25" Cyclops requirement

for the X and Y axes, and under 0.25" from the calculated ISIPOD force vector, thereby

satisfying the CG location requirements.

28
 AGS4 was designed as a series of panels that would fold up into the completed

box structure. This approach was helpful as it divided the spacecraft up into regions

which was an easy way to designate component locations. A custom assembly jig was

developed to keep all the panels in the appropriate relative position and orientation such

that wires could be run across the panel gaps to supply power and data to the

components on that face. Once all parts were integrated into their panel, the panels

would be folded in and attached to one another. The figures below show how the panels

were laid out for assembly.

Figure 9: Panel Assembly Configuration

29
 Figure 10: AGS4 Assembly Jig and Panel Layout

All structural materials on the vehicle were made out of 6061-T6 aluminum.

This aluminum grade was selected for its weight, and its common use in aerospace

applications. As the structural analysis indicated that this material was strong enough to

endure the expected loads, the stronger but heavier 7075 series aluminum was not

needed. In order to preserve grounding and bonding between all structural elements

except the solar panel trusses, Alodine was applied to all bonding surfaces. Alodine

is a corrosion prevention treatment that maintains electrical connectivity. The remainder

of the aluminum surfaces were anodized with Type-II soft coat to prevent electrical

conductance in places not desired. Soft coat anodization was chosen over Type-III hard

30
coat as it is a much simpler process, and the surfaces of AGS4 are not expected to

experience high amounts of friction or wear over its lifetime.

Locking inserts are necessary to prevent fasteners from backing out due to the

launch vibrations, which could cause pieces to fly apart [15]. There are many approved

methods to prevent fastener back-out, but the one favored by the AGS4 team was

locking threaded inserts from Heli-Coil. To use Heli-Coils, the fastener hole is

drilled to a prescribed diameter that is larger than would be for the fastener alone. The

hole is tapped with a specific tap and die set, and the Heli-Coil is inserted using the

installation tool, similar to a fastener installation. The Heli-Coil has specially designed

surfaces that cut into the hole material to lock it in place, and similar surfaces on the

inside to cut into and lock onto the inserted fastener, preventing back-out. Also recently

approved for use was Locktite epoxy, but this was only used on the #0 size fasteners

since Heli-Coils are not made that small. All of the Heli-Coils were installed by the

AGSL students, as having the machine shop install the hundreds of them would have

been costly. The process is not difficult, and previous student experience was leveraged

to teach the other lab members how to perform the procedure.

2.4.1.2. Other Structures

In addition to the bus, structural elements were designed for the backing of the

solar panels, to prevent them from flexing and cracking the solar cells. Frequently

aluminum honeycomb panels are used for this purpose as they are extremely rigid and

unbelievably light. Unfortunately this option would produce panels that were too thick

and would have clearance issues with the airlock, and so rather than redesign the whole
31
bus to shrink it down and accommodate the thick panels, AGSL designed thin,

lightweight aluminum trusses to support the panels.

Figure 11: Solar Panel Truss Structure (inches)

Figure 12: T-Shaped Solar Panel Truss (inches)

32
 Figure 13: Long Solar Panel Truss (inches)

The solar panels were affixed to these structures using epoxy, and attached to the bus

using screws.

Similar to the solar panel structures, the panels required to have some sort of

cover to protect them from impact and contain any potential glass fragments that might

come loose into the ISS cabin environment [16]. In the micro gravity environment, tiny

glass particles will find their way into sensitive equipment or possibly astronauts’ eyes.

Covers would need to be able to withstand a 25 lb distributed force to protect against

accidental astronaut impact, and would need to be see-through so that the cells could be

inspected for cracks or breaks before removal. If a crack were to be discovered, the

cover for that panel would be left in place to prevent cabin contamination.

The initial thought was to use polycarbonate sheet as the cover since it is strong

enough to withstand the impact, clear to see through, and relatively lightweight.

Unfortunately, there was not enough room in the mass budget to fly this design, so the

team had to get creative. Instead of a solid sheet, a waterjet was used to cut triangular

33
holes in the sheet to retain strength but lose weight. Another polycarbonate sheet would

be cut to provide a 1/4" standoff between the grid sheet and the cells. A Mylar plastic

film was epoxied to the inside of the grid to retain visibility but provide containment.

The result was a cover with a mass of just 122 g. Velcro strips were epoxied to the

cover frame, and a mating piece was epoxied to the solar panel/truss combination in

order to secure the cover. Astronauts would pull on these Velcro tabs to remove the

covers. The figure below shows the finished square panel covers.

Figure 14: Solar Panel Cover

34
 There was not enough clearance on the bottom of the spacecraft for it to be

attached to the Cyclops table if any of the cells broke and required the cover to stay on.

To prevent contamination risk, Mylar covers were cut for each of the bottom panels

and held in place with Kapton tape. This would provide no impact protection, but

would provide containment. If all cells were intact after install onto Cyclops, the tape

tabs could be pulled and the Mylar slid off.

Figure 15: Bottom Panel Mylar Covers

One other item that was developed, and that AGSL was a pioneer of, was the use

of 3D printed materials in spacecraft. After learning that polycarbonate was an

35
acceptable material for both inside and outside ISS from a temperature and outgassing

standpoint, it was decided to be used for brackets to secure wires. AGS2 had used

Teflon for a similar purpose, but this required expensive material and machining. The

TAMU Aerospace Engineering department machine shop had high end 3D printers that

could print in polycarbonate, and thus AGSL was able to create custom wire brackets for

less than a dollar a piece. These brackets not only were of the correct custom size and so

therefore did not require shimming to prevent wire wiggle, but they do not have sharp

edges like metal brackets that require covering. Some of these brackets can be seen in

the above image in the lower left quadrant, the white blocks with the wires passing

through them. They can also be seen clearly in the figure below.

Figure 16: Polycarbonate Wire Brackets

36
 Figure 17: Polycarbonate Wire Bracket Model

2.4.1.3. Wire Modeling

All of the component wires were modeled in SolidWorks. While a tedious and

difficult process, it was necessary in order to determine the correct length for each wire,

and the correct size for each wire bracket. Proper bend radius for the gauge wires used

was implemented, which also determined the paths that the wires would take from one

component to another. Wire "corridors" began to develop, where all the wires going in a

general direction were clustered together to keep them organized. This modeling also

served to more accurately model the mass and center of gravity of AGS4.

2.4.1.4. Foam Packaging Enclosure

In order to protect AGS4 as soft stow cargo for the launch, NASA cut and

assembled approximately 2" thick LD45FR foam padding in the form of two clamshell

halves. This was based on the AGS4 exterior CAD model. The foam was designed to

37
fit snugly, prevent damage from impacts, and dampen the vibrations from the launch

vehicle.

Figure 18: Clamshell Foam Lower Half

38
 Figure 19: Clamshell Foam Upper Half

Figure 20: Foam Packaging Housing AGS4

39
2.4.1.5. Structural Analysis

In order to ensure that the design would survive the acceleration and vibration

loads of the launch vehicle, a structural analysis was performed [17]. Using

SolidWorks, gravitational loads were applied to mirror the a Falcon-9's acceleration,

and a dynamic random vibration power spectral density (PSD) curve was applied to

mirror the vibrations from the engines.

Before any analysis can begin, the boundary conditions must be defined. These

were derived from the method by which AGS4 would be strapped to the launch vehicle

inside the foam packaging. The figure below details how the package would be attached

to the vibration testing table and the launch vehicle.

Figure 21: AGS4 Launch Restraint Configuration (Image by NASA)

40
 As the foam packaging was designed to provide open space for the solar panels

and other features to not experience any contact, the weight of the structure was focused

on a few structural points. These points were modeled as fixed geometry in

SolidWorks, and are at the bus corners and the Cyclops spacer as shown in the figures

below.

Figure 22: Top Fixed Constraints

Figure 23: Bottom Fixed Constraints

41
 Without the resources of a supercomputer, modeling the entire model to the

highest degree of fidelity would not be possible. Simplification to the components was

needed to cut down the model complexity. Electronics boards were simplified down to

just the FR4 plastic boards with the appropriate density to account for the surface

components that were suppressed. The ISIPOD model was simplified, and Bevo-2

replaced with a simple block of the same mass. Wires and fasteners were omitted as

their complexity was far too great, and they will likely experience significant amounts of

low amplitude oscillation between mounting brackets which would make simulation

impossible with the high end computers available to the lab.

SolidWorks uses finite element method (FEM) analysis to determine the

stresses and deflection a structure experiences under load. FEM divides the model into a

mesh of roughly equal size pieces, and uses the conditions of the neighboring pieces as

the boundary conditions for the piece in question. Structural stress and deformation

calculations are performed on the piece in question. Then, the piece that was just

examined is used as the boundary condition for the next piece in its new deformation

state. This process is continued across the entire mesh and repeated using the results of

each piece in the next iteration until convergence is reached, and the results do not

change significantly from one iteration to the next.

With the simplified model developed, the next step was to determine the

appropriate mesh size to perform the analyses. Too coarse an analysis will yield

inaccurate results, while too fine a mesh will spend orders of magnitude more time for

an unimproved result. A convergence analysis was performed using the SolidWorks

42
natural frequency calculation function. The analyses started with coarse meshes and

progressively increased the number of nodes in the system until the 1st mode natural

frequency did not change significantly with more nodes. As shown in the table of results

below, this occurred at a 1.5 cm mesh.

Table 7: Frequency Modes for Mesh Sizes

Mesh Size 2.5 2.25 2 1.75 1.5 1.25 1
(cm)
# of Model 238032 277943 306549 352220 433309 552024 762143
Nodes
Frequency 311.08 307.08 301.86 298.26 288.53 284.21 281.1
Mode #1
(Hz)
Frequency 317.66 315.93 313.15 310.59 303.87 299.97 297.63
Mode #2
(Hz)
Frequency 335.44 332.1 324.68 319.38 309.57 305.53 303.22
Mode #3
(Hz)
Frequency 378.41 373.45 370.91 365.17 355.76 350.53 351.89
Mode #4
(Hz)
Frequency 398.30 395.47 394.38 392.06 383.73 378.26 371.55
Mode #5
(Hz)
Run Time 00:08:32 00:09:24 00:10:04 00:12:04 00:40:28 00:51:01 01:13:48

43
 Figure 24: 1.5 cm Analysis Mesh

Since the 1st mode natural frequency is above 280 Hz, it is not expected to

experience resonance with the launch vehicle that would induce structural failure. A

natural frequency above 100 Hz is desired to prevent resonance. The items that are most

susceptible to vibrations are the electronics boards.

44
 Figure 25: 1st Mode Natural Frequency Response

In order to perform the random vibration analysis, the launch environment needs

to be known. Since at the time of the analysis it was unsure whether OA-4 would fly on

an Atlas V or Falcon-9 rocket, the NASA management had AGSL perform the analysis

with the Falcon-9 environment as this was the more harsh of the two [16]. The random

vibration environment of the vehicle is outlined in the table below, and the PSD curve in

the following figure.

Table 8: Falcon-9 Random Vibration Environment

Frequency (Hz) Level (g2/Hz)
20 0.057
153 0.057
190 0.099
250 0.099
750 0.055
2000 0.018
Composite 9.47g rms
Duration 60 seconds/axis

45
 SpaceX Dragon Random Vibration Environment
1.00000

0.10000
g2/Hz
PSD

0.01000

0.00100

0.00010
10.000 100.000 1000.000 10000.000
Frequency
Hz

Figure 26: Falcon-9 PSD Function

However, in order to determine that the payload is safe to experience these loads,

NASA tests to +3 dB, or double the power. Fortunately for soft stowed payloads like

AGS4, the foam packaging attenuates a great deal of this vibration power. Since foam

can have non-linear properties, including it in the simulations would increase the

complexity by orders of magnitude beyond AGSL computing capabilities. Fortunately,

NASA provided the "Cargo Tool" [18] which utilizes the surface area of the spacecraft in

contact with the foam to determine the attenuation so that it can be left out by testing with an

adjusted curve. Two inches of LD45FR foam was used in the Cargo Tool to produce the

attenuated +3 dB PSD curve shown in the figure below. The 0.7 psi curve applies to the X

and Y axes, and the 0.2 psi curve applies to the Z axis. These psi values are determined

46
based on the mass of the spacecraft and the surface area of foam in contact with the faces on

that axis to distribute the weight.

Figure 27: Attenuated +3 dB PSD Function

The analysis was run using these curves for each axis. This produced a

calculation of maximum instantaneous acceleration and stress at each point in the model.

The results calculated were that the lowest factor of safety anywhere in the model under

these loads was 565 at the location that the Cyclops attachment fixture attaches to the

standoff spacer. This is a good indicator that none of the materials will fracture due to

the vibration environment. Unfortunately, there are things that cannot be properly

modeled such as wiring, soldered connections, and integrated circuit silicon wafers

which can break due to vibrations, and while they do not pose an imminent safety

47
hazard, they do harm the ability to meet the mission objectives. Adherence to

workmanship standards [15] and proper soldering techniques [19] reduce this risk.

This same analysis was applied to the solar panels (with structural truss backing)

and solar panel covers. As the bottom face panels do not have structural covers, they

were analyzed as just the panels with their truss backing.

Figure 28: Solar Panel Vibration Analysis

The results of this analysis indicate that these structures are not at risk of

fracturing due to vibration loads.

Table 9 : Solar Panel Vibration Factor of Safety

Panel 1 sigma acceleration Stress (Pa) Factor of Safety
(g)
T-Shaped 2.2596 11567 23775
Long 2.1767 7424.4 37040
Square with Cover 2.711 19966 13773
(X/Y-Axis)
Square with Cover 7.074 119599.4 1438
(Z-Axis)

48
 With confidence that the vehicle and components will not fracture due to

vibrations and resonance, the analysis of stresses and deflections due to acceleration are

analyzed in a similar fashion. SSP 50835 [16] states that the maximum acceleration load

that the SpaceX Dragon is capable of experiencing is 9.0 g's, though actual loads are

much less, and that payloads should be analyzed to this value in all axes. Additionally,

payloads should be able to withstand rotational accelerations of the launch vehicle of up

to 13.5 rad/sec2. The specification also requires that all payloads have a factor of safety

of at least two.

This analysis was setup the same way as the vibration analysis. An acceleration

of 9.0 g was applied to each axis, both positive and negative directions, while a

rotational acceleration equal to 13.5 rad/sec2 was applied as if the payload were mounted

to the outside wall 1.8 m away from the center. This was to simulate any of the six

spacecraft faces being in the vertical position during launch, while the rocket performs

the rotation maneuver. The resulting maximum displacements and minimum factor of

safety per axis are detailed in the table below.

Table 10: Axial and Rotational Acceleration Results

Factor of Safety Displacement (mm) Max. Displacement
Location
X-axis 4.04 0.02768 EPS Boards
Assembly
Y-axis 3.3 0.04522 EPS Boards
Assembly
Z-axis 2.38 0.04277 CDH Assembly

49
 Figure 29: +Z Axis and Rotational Acceleration Displacement

This same analysis was also applied to the solar panel structures and solar panel

covers. Under their own weight due to the accelerations, as expected the highest stress

and displacements of the panels occurred when they were oriented perpendicular to the

axis of acceleration.

Figure 30: T-Shaped Panel Displacement

50
Figure 31: Long Panel Displacement

Figure 32: Square Panel Displacement

51
 Figure 33: Panel Cover Displacement

Table 11: Panel Acceleration Stress and Displacement

Minimum Factor of Safety Maximum Displacement
(mm)
T-Shaped Panel 7.70 0.04466
Long Panel 53.45 0.04096
Square Panel 19.29 0.06057
Panel Cover 2.63 5.624

Since all factor of safety requirements were satisfied, the next concern was that

the maximum displacement would be enough to crack or damage the solar cells fixed to

them, or that the panel covers would deflect far enough that they would make contact. A

simple test was done on a sample solar cell and it was found that the cell could flex by

roughly 1 mm without cracking the cover glass, more than enough to tolerate these

displacements. The solar panel covers on the other hand were a much closer call. The

grid structure sits 1/4", or 6.35 mm off the face of the solar cells. Given the maximum

displacement of 5.6 mm, there is only a small amount of margin under maximum load.
52
Still, this was sufficient to protect the cells, and maximum loads were unlikely to be

experienced.

This structural analysis indicated that the AGS4 configuration would survive

launch and reach the ISS as a viable payload, without posing a hazard to the crew or

station.

2.4.1.6. Thermal Analysis

A thermal analysis [20] was required to verify that AGS4 could withstand being

outside the ISS for up to 10 hours while coming out of the airlock and being manipulated

by the robotic arm. Unlike the structural analysis, which could leave out components

such as the solar panels which do not significantly contribute to the dynamic response, a

thermal analysis requires a higher fidelity simulation model which includes these items.

This is because these components provide surface shading, which impacts internal

temperatures due to their impact on radiant heating. Fortunately including these

features does not pose quite as much a computation burden as it does on the structural

analysis, since a great deal of the heat transfer is by internal conduction, which results in

a larger mesh size being adequate to calculate temperatures. Using the SolidWorks

thermal analysis package, a test case was run as a convergence analysis, and it was found

that a mesh of 5 cm is sufficient to accurately model the thermal properties.

53
 Figure 34: AGS4 Thermal Simulation Model

Figure 35: 5 cm Mesh

In a purely radiation heat transfer environment like the vacuum of space, the

amount of energy an object absorbs is related to the view factor of the object’s face to

54
the heat source. The view factor is a percentage of the face that is exposed to that

source. Assuming a 400 km orbit and that the X+ panel is always facing Earth to

simulate the ISS Local Vertical Local Horizontal hold (i.e. rotating to keep one side of

the ISS constantly pointed at Earth), the six panels have view factors according to the

table below.

Table 12: AGS4 ISS View Factors

Panel Earth View Factor Free Space View Factor
+X 93% 7%
+Y 22% 78%
+Z 22% 78%
-X 0% 100%
-Y 22% 78%
-Z 22% 78%

The view factors are significant in that they determine the radiation heat flux to

or from that face. There are three primary sources of heat flux in the low Earth orbit

(LEO) environment: solar heat flux, outgoing long-wave radiation (Earth blackbody),

and Earth albedo (reflected sunlight). These radiation sources have power densities as

shown in the table below. Also shown are the Earth and free space temperatures which

heated surfaces will radiate to according to their view factor and temperature differences.

Table 13: Radiative Heat Flux Sources and Sinks

Solar Flux 1367 W/m2
Outgoing Long-wave 287 W/m2
Radiation
Earth Albedo 410.1 W/m2
Free Space Temperature 2.725 K
Earth Temperature 255 K

55
 Using the above parameters, AGS4 was evaluated over the course of 10.5 hours

for three different beta angle cases, 0°, ±60°, and ±75°. The beta angle is the angle

between the spacecraft orbit plane and the Earth-Sun vector, which changes with the

seasons. The case of β=0° is the coldest scenario where AGS4 spends 59% of the time

in the sun, while β=75° is the hottest case when the orbit is 100% illuminated by

sunlight.

SolidWorks uses radiative heat transfer mechanics to evaluate the external heat

flux, and conduction and surface-to-surface radiation to determine the temperatures of

each spacecraft component. Each component was monitored to ensure that its

temperature stayed within appropriate bounds during the simulations. The primary

concern was the lithium ion battery cells which had the smallest temperature range of -

20-60°C. The batteries came the closest to their temperature limit of all the components,

but stayed within bounds for all three Beta angles. Over the 10.5 hour simulations, they

reached a minimum -9.7°C at β=0°, and a maximum of 27.4°C at β=75°.

56
 Figure 36: Temperature Model for β= 75°

This analysis indicates that AGS4 will be able to survive the up to 10 hours in an

unpowered state while in close proximity to the ISS. Since the orbit will be very similar

in shape and altitude for the majority of the spacecraft lifetime, this analysis also

indicates that AGS4 should not experience temperature problems during operations.

2.4.2. Command and Data Handling (CDH)

The CDH system is the main computer on the spacecraft, and is responsible for

facilitating communications among all subsystems, as well as executing the mission

script and performing calculations. AGSL had manufactured the CDH system in-house

based around a microprocessor for the less complicated AGS2 mission. Experience

from this mission recommended that AGSL upgrade to single board computers for future

missions [21] as their more conventional structure and software would ease development

and operations.
57
 During the trade study process, there was some trouble finding computer systems

that were of the appropriate size and computing power that had sufficient input/output

ports to interface with all of the components. CubeSat components usually use the

PC104 format which connects stacks of boards with 60+ pin connectors, and do not have

ports for traditional components. After several rounds of investigation and trade study,

the CDH team found the TS-7800 single board ARM computer from Technologic

Systems [22].

Figure 37: TS-7800 Flight Computer

The TS-7800 processor operates at 500 MHz, has 128 MB RAM, and is designed

to operate in rough environments. Of particular interest was the computer's ability to

accept TS-SER4 serial expansion boards, providing enough ports for all of the AGS4

components. This feature, in addition to the two USB ports needed for the Kenwood

radios, and the SD card slot for easy upload and storage of the flight software, made the

TS-7800 a computing option that supported the mission and all components.

58
 Figure 38: TS-SER4 Serial Expansion Board

A Linux operating system was chosen for its flexibility and customizability to be

tailored to the AGS4 mission, and its suitability to work on ARM single board

computers.

Figure 39: CDH Stack

59
2.4.3. Electrical Power System (EPS)

After performing a rough-order-of-magnitude analysis for the power

requirements of AGS4, it quickly became apparent that commercially available CubeSat

power systems of the day were insufficient to meet the mission needs. Since AGSL had

team members with the skill set and previous experience with developing small power

systems, the choice was made to fabricate custom designed EPS hardware.

Table 14: AGS4 Power Budget

Voltage Current Power
Subsystem Device (V) (A) (W) Quantity

ADCS Vector-Nav 5.00 0.07 0.33 1

ADCS Reaction Wheel 28.00 0.25 7.00 3
ADCS Torque Coil 28.00 0.20 5.60 3
TS-7800 Single Board
CDH Computer 5.00 0.80 4.00 1
CDH TS-SER4 Serial Board 5.00 0.08 0.40 3
5.00 0.73 3.65
COMM Digi Xtend TX/RX 5.00 0.08 0.40 1
Kenwood TH-D72A 9600 12.00 1.84 22.08
COMM TX/RX 12.00 0.12 1.38 1
Kenwood TH-D72A 1200 12.00 1.84 22.08
COMM TX/RX 12.00 0.12 1.38 1
High Data Rate – Radio
Frequency IC (HDR –
COMM RFIC) 5.00 0.02 0.11 1
High Data Rate –
COMM Amplifier (HDR – AMP) 12.00 1.60 19.20 1
DRAGON DRAGON Board 5.00 0.30 1.50 2
SMTRS ISIS ISIPOD Deployer 28.00 1.75 49.00 1
VDCS PSCAM 5.00 1.00 5.00 1
Bevo Charging Bevo-2 5.00 0.50 2.50 1

60
 Figure 40: AGS4 Power System Diagram

Figure 41: EPS and Battery Installed on AGS4

61
 The EPS consisted of the following custom designed and manufactured

components:

2.4.3.1. Battery Charge Regulator

The battery charge regulator board operated as the heart of EPS. In addition to

ingesting power from the solar cells, charging the batteries, and distributing solar/battery

power to the other EPS stack boards, this board contained a microcontroller to direct all

EPS functions across all boards.

Figure 42: Battery Charge Regulator Board

2.4.3.2. Voltage Regulator Board

The voltage regulator board used a series of regulation circuits and inductors to

change solar panel/battery voltage into the various levels the output boards and

components require.

62
 Figure 43: Voltage Regulator Board

2.4.3.3. Torque Coil Control Board

The torque coil control board controlled the 28 V current supplied to each of the

three torque coils used for momentum desaturation and coarse pointing of AGS4.

Figure 44: Torque Coil Control Board Layout

63
2.4.3.4. Output Board

The output board would turn on and off power at specific ports based on

commands from CDH. The different voltage levels were produced by the voltage

regulator board. The board is capable of being configured for output at 5V, 12V, and

28V. Three of these boards were used to provide the necessary number of output ports

at the proper voltage to power the equipment as per the power budget.

Figure 45: Output Board

2.4.3.5. Hazard Control Board

The hazard control board was designed to prevent CDH from commanding, and

EPS from being able to supply power to the ISIPOD release mechanism until a NASA

determined time had elapsed. This was to ensure that the release of Bevo-2 could not

come back and hit the ISS.

64
 In order to perform this function, this board carried three separate timers powered

by independent coin-cell batteries, which would start counting after the inhibits were

released. Until all three had shown to have exceeded the predetermined time allotment,

relays would remain in the open position physically preventing power from flowing to

the release mechanism. The timers would all count up instead of down so that in the

event of a power outage they had to start all over rather than immediately being at zero.

Figure 46: Hazard Control Board

2.4.3.6. EPS Controller Board

The EPS controller board was the means of digital interface between the EPS

stack, primarily the battery charge regulator board, and CDH. Control signals for

various components are received from CDH, and acted upon by the EPS microprocessor.

65
 Figure 47: EPS Controller Board

2.4.3.7. Battery and Protection Circuit

As available CubeSat battery packs were not sufficiently large enough for the

needs of AGS4, and batteries for larger spacecraft came with a hefty price tag, the

decision was made to develop custom battery packs for AGS4. AGSL had gained

experience developing small batteries for AGS2, and that experience would be leveraged

along with team members experience building Lithium Ion (Li-ion) packs to produce

batteries to meet the spacecraft power requirements.

AGS4 would carry two battery packs consisting of nine LG ICR18650C2 Li-ion

cells, each with 2800 mAh capacity [23]. The 18650 cell format was used by AGSL on

AGS2 for its low mass and high energy density. Nominal voltage was 3.75 V, and when

wired in series the pack voltage was 33.75 V.

66
 Figure 48: Battery Prototype

As battery explosions are a very real possibility with Li-ion batteries, steps were

taken to contain any explosion and protect the rest of the spacecraft, and more

importantly the ISS. The nine cells were housed in roughly 3.5x3.5 in. anodized

aluminum boxes with divider pockets for each cell. The anodization is intended to

passivate the aluminum surfaces, preventing unintended conduction through the box

structure, and the box itself is intended to keep any battery explosions contained.

67
 Figure 49: Battery Cut-Away View

The box was lined with, and each cell was wrapped with sheets of Durette felt,

which is a filter material that provides thermal insulation as well as filtration of any

liquids that may be released in the event of an explosion. In order to prevent the battery

box from becoming a pressurized vessel, whether from atmospheric pressure change or

from battery explosion, vent holes were included in one side of the box. These holes

were covered with layers of Durette felt to filter the gas escaping.

68
 Figure 50: Battery Box Vent and Cover

Teflon insulated 12 ga. wire was soldered to each battery terminal to connect

them in series. The wires were routed through the upper gaps in the aluminum dividers,

and then they exited the box through two 1/4 in. holes in the lid and ran to the battery

protection circuit. Smaller gauge wires were also attached to the positive terminal of

each cell, and terminated in a micro DB-9 connector, also in the lid.

The battery protection circuit was necessary to ensure that no individual cell was

overcharged or undercharged, conditions that could have destructive results. The circuit

was based on the Texas Instruments BQ77PL900 circuit, which used the smaller wires

attached to each positive terminal to monitor each cell’s voltage, and balance any cells

that were over/under voltage from the others.

69
Figure 51: Battery Protection Circuit

Figure 52: Assembled Battery Box

70
2.4.3.8. Solar Arrays

The AGS4 team custom designed all of the solar panels using printed circuit

boards (PCB) and Spectrolab UTJ solar cells [24]. Custom panels were chosen because

body fixed panels can be tailored to the specifics of the vehicle for little cost, and

because the offerings on the market were either too large or too small for a spacecraft of

this size. Spectrolab UTJ cells were selected as they were one of the highest efficiency

cells available at that time, and could be procured in the needed timeframe.

AGS4 had two 6x3 solar panels on each of its faces, except for the bottom face

which had three odd sized panels. The 10 panels on the top and side faces were made

from white PCB's in order to reduce the amount they would heat up in the sunlight.

Figure 53: Top and Side Solar Panel Configuration

71
 The bottom face had two 3x6 panels and one T-shaped panel containing 18 cells.

These panels were shaped this way to accommodate the Cyclops hardware. These

panels were green in color because the manufacturer that could print PCB's in this size

could only do so in green. All panels on the vehicle had 18 cells to ensure even power

generation capability.

AGSL was one of the first spacecraft developers to attach solar cells to the panels

using double sided Kapton tape, instead of the traditional epoxy. This process will be

described subsequently in the assembly section.

Figure 54: Bottom Face Solar Panel Configuration

2.4.3.9. Wiring Harnesses

All power and data wire harnesses were custom designed. While some were

completely handmade to specification using Teflon coated wire and connectors, most
72
were custom made by Omnetics, a well respected manufacturer of spaceflight wire

harnesses. Some of the Omnetics harnesses were purchased with connections on one

end only, so that AGSL could customize the other end to the specific hardware need.

Appendix B contains an example of the Omnetics harness specification for a reaction

wheel.

2.4.3.10. Inhibits

In order to prevent AGS4 from prematurely activating and potentially releasing

BEVO-2 while inside the ISS or prior to release from the Cyclops table, NASA required

two sets of three-fault tolerant inhibits.

The first inhibit set was an activation key made out of a Teflon backed DB-9

connector. Current from the solar panels had to flow through four sets of pins in the

connector in order for EPS to be able to activate. Without this key in place, the physical

connection is broken preventing current from flowing. When AGS4 was fully installed

on the Cyclops table and prepared to enter the airlock, the astronauts removed the key

from its "SAFE" storage position, and installed it in the "ACTIVE" receptacle using the

screws to prevent it from coming loose.

73
 Figure 55: Inhibit "Active"

In order to keep AGS4 inhibited once the key was moved to "ACTIVE", AGS4

contained four linear pressure switches through which the key current had to flow.

When installed on Cyclops, these switches were depressed and held on the open

position, preventing current flow. Once released from Cyclops, the switches would

depress and allow the activating solar panel current to flow to EPS. The switches were

three fault tolerant as they were wired in series so that even if three switches failed to

break the circuit, one would still prevent activation.

These four switches were located in the empty space on the inside of the Cyclops

knob. They were located here as this was one of the few places where they could be

assured to make contact with the Cyclops table and remain depressed.
74
 Figure 56: AGS4 Inhibit Switches

Originally, AGS4 had been designed with four small feet roughly 8 in. outside

the knob, and these feet were to house the switches. However, as the Cyclops system

was developed in parallel with AGS4, design changes on that system necessitated

moving the switches inboard to prevent the possibility of the switch force inducing a

spin at release.

75
 Figure 57: Prior AGS4 Foot Switch Design

Figure 58: Previous Design Switch Preparation

76
2.4.4. Communications (COMM)

Lessons learned from the communications troubles on AGS2 were implemented

in the communications systems of AGS4. Frequency hopping spread spectrum (FHSS)

was avoided for space to ground communications due to its problems compensating for

Doppler shift [21]. Instead, it was decided to use redundant lower data rate frequency

modulation (FM) radios for simplicity and redundancy.

For reliability, the system design would consist of separate bi-directional uplink

and downlink radios. Nominally each would operate only one direction, but had the

capability to serve both uplink and downlink roles if needed. Additionally, AGS4 would

carry an experimental high data rate transmitter to bring mission data down more quickly

if able. Finally, AGS4 would carry a separate crosslink radio to communicate with

Bevo-2.

2.4.4.1. Trade Study Criteria

To determine if the candidate hardware was sufficient to meet the mission

requirements, various trade studies were performed for each and compared. Factors

compared were:

1. Link Budgets: A link budget analysis determines if the signal strength of

the transmitters is sufficient to transmit the desired data rate given the

receiver equipment, ambient noise level, modulation scheme, etc. Details

of link budget calculations are covered in a subsequent section.

2. Environment: All equipment must be rated by the manufacturer to

operate within the bounds of the expected environmental conditions,

77
 largely temperature and radiation. As LEO is relatively warm and low

radiation, and given the relatively short life expectancy of AGS4, it was

determined that commercial grade equipment would be sufficient.

Pockets of radiation such as the South Atlantic Anomaly were not

understood or accounted for in this analysis, and should be for future

efforts before proceeding with this grade hardware.

3. Data Budget: A calculation of data generated over time and mission phase

determines how much throughput the communications system will

require. By using orbit tracking software, an expected number of

communications passes per week can be estimated. Using data from

AGS2 regarding personnel availability to monitor passes, weather

conditions, and other issues that impeded communications, an adjusted

estimate of data per week downloaded was calculated and used to justify

the hardware data rates.

4. Cost: Cost of the units compared to capability is a factor in selecting

hardware.

5. Frequency and Licensing: Choosing a frequency which has a wide variety

of equipment available to operate in that band provides hardware options.

Also, selecting a frequency band for which a license will be easily

granted reduces complexity and paperwork. AGSL chose to apply for an

experimental FCC license using amateur radio frequencies. Experimental

licenses are granted for projects to develop hardware and software

78
 capability such as with AGS4, and using amateur frequencies allows

radio operators around the world to track the beacons and potentially

perform ground passes.

By analyzing the candidate components along these criteria, AGSL converged on

a system design of components outlined in the subsequent sections.

2.4.4.2. Data Budget

In order to determine the necessary radio bandwidth to downlink all the required

mission data, an accounting of all of the data generated by the spacecraft must be

performed. Spacecraft data falls into three main categories: health data, payload data,

and subsystem data. Since data transmission is typically in rates of bits and megabits

(Mb), this standard is used in the data volume generation calculations as opposed to the

traditional storage units of bytes and megabytes (MB).

Health data is generally limited to status data about the spacecraft or subsystem

components. These values are constantly monitored by CDH to keep them in the proper

bounds, and in the case of AGS4 were stored in memory once every five minutes. With

this five minute interval rate, a fairly granular time history of AGS4 could be

reconstructed over the elapsed orbits. The state of health (SOH) data packet is listed in

the table below, outlining what bits of information were monitored and stored.

79
 Table 15: AGS4 State of Health Packet
State of Health
Bit Size Qty. Subtotal
Battery Pack Voltage 16 2 32
Battery Pack Current 16 2 32
Battery Pack Temp 16 2 32
Charge Level Indication 16 2 32
Battery Charge Status 2 2 4
EPS Board Temps 16 5 80
Solar Panel Current 16 6 96
Safety 8 2 16
State of Components 8 2 16
CDH Current 16 1 16
CDH Temp 16 3 48
Timestamp 64 1 64
Remaining Data Storage 64 1 64
Data Remaining to Downlink 64 1 64
Current Program mode 8 1 8
COMM Current 16 7 112
COMM Temp 16 7 112
Acceleration (x,y,z) 32 3 96
Angular Velocity (x,y,z) 32 3 96
Attitude Representation 32 4 128
Magnetometer Reading 32 3 96
ADCS Current 16 7 112
ADCS Temp 16 7 112
DRAGON Current 16 2 32
GPS Packet 6F (internal health reading) 272 2 544
Pos/Vel. Solution (GPS DATA) 384 1 384
SMTRS Temp 16 6 96
VDCS Current 16 1 16
VDCS Temp 16 1 16
ISIS Current 16 1 16
Bevo-2 Current 16 1 16
Third Party Current 16 7 112
Third Party temp 16 7 112
2588 bits/sample
80
 At the rate of one sample per five minutes, the total estimated SOH data is

calculated in the table below.

Table 16: Health Data Volume

Number of SOH Samples
Estimated Mission Lifetime (shortest) 3 months
Sampling Rate once per 5 minutes

Samples 25920
Total State of Health Data 72.89 Mb

The payloads for AGS4 were the Dual RF Astrodynamic GPS Orbital Navigator

(or DRAGON) GPS receiver and the camera system, both described in subsequent

sections. The mission requirements call for two orbits of raw GPS data as per MSC-8,

and AGSL initially chose one orbit of carrier phase data to satisfy the feed forward

requirement of MR-9. This GPS data generated is outlined in the table below. The

carrier phase data would be collected during the second raw GPS orbit to utilize data that

was already requiring downlink.

Table 17: DRAGON GPS Data Generated

DRAGON Data
Two Orbits Raw GPS Data, MSC-8 52.10 Mb
Crosslinked Solutions, MSC-10 0.80 Mb
One Orbit Carrier Phase Data, MR-9 78.6 Mb
GPS Data, 8 hr. Tracking 10.1 Mb

DRAGON Download 141.60 Mb

The mission requirements call for there to be visual confirmation of the AGS4

and Bevo-2 separation event. AGS4 would carry a camera, described in a subsequent

section, mounted by the ISIPOD to record the event. This camera would operate in a

81
medium resolution burst mode to capture the separation, and a slower speed full

resolution mode for the Earth imaging phase. Thumbnails of images would be

downlinked to reduce data volume, and the best examples selected for full resolution

downlink. The table below outlines the image sizes and data generated, though they are

likely to be lower with JPEG compression applied.

Table 18: Visual Data Downlink Volume

VDCS Data
Full Resolution Full Resolution Image size 46.1 Mb/image
Video Video Frame Size 7.4 Mb/image
Thumbnail Thumbnail size 0.26 Mb/image

Release Event Duration of Video 40.00 sec

Frame Rate 1.00 frames/sec
Release Video Image
Quantity 40.00 images
Size of Download 296.00 Mb
Earth Imaging Duration of Phase 90 sec
Frame Rate 0.25 Frames/sec
Earth Image Quantity 23.00 images
Full Resolution to
download 3.00 images
Thumbnails to download 23.00 images
Size of Download 144.28 Mb
Total VDCS Download 440 Mb

The only AGS4 subsystem that required data recording outside of the state of

health readings was the attitude determination and control system (ADCS). This

subsystem data was required to verify the spacecraft orientation and control during the

crosslink experiment, GPS data collection phase, and Earth imaging phase. The ADCS

data packets contain the following information and data sizes.

82
 Table 19: ADCS Data Packet Size
ADCS Downlink Data

Component Bit Size Qty. Subtotal

Torque Coil (AGSL Manufactured)
Temperature 16 3 48
Current 16 3 48
Reaction Wheel
Temperature 16 3 48
Current 16 3 48
RPM 16 3 48
VectorNav (VN-100T)
3-axis gyros 32 3 96
3-axis accelerometer 32 3 96
3-axis digital compass (magnetometer) 32 3 96
Temperature 16 1 16
Total ADCS Hardware Sample (Bits Per
Sample) 544

Using this data rate, the ADCS data for each of the mission phases can be totaled

as in the table below.

Table 20: Total ADCS Data Generation

ADCS
GPS Data Logging (2-orbits) 2.94 Mb
Bevo-2 Release 0.05 Mb
Bevo-2 Tracking (8 hr.) 15.7 Mb
Earth Imaging 0.05 Mb

Total ADCS Data 18.74 Mb

With the data generated for health, payloads, and subsystems totaled, the grand

total data required to be downlinked is the sum of these three categories.

83
 Table 21: Total Downlink Data
Total Data Notes
Raw GPS data, RelNav exchanged data,
DRAGON 141.60 Mb Carrier Phase Data, Bevo-2 tracking
Health 72.89 Mb every 5 min for life
VDCS 440.00 Mb
tracking, stabilization, ground pass,
ADCS 18.74 Mb imaging

Total 673.23 Mb

This total downlink data calculation is then used to determine what data rates are

sufficient for the downlink radio system. As is further defined in the subsequent

sections, two downlink systems are included on AGS4, and the data rates under scrutiny

are 9600 bits/second and 153.6 kilobits/second.

Unfortunately, it is not as simple as just dividing the total data by the data rate.

Most of the time spacecraft are not within range of the ground station, and even when

they are, the highest possible data rate is rarely achieved. In order to determine the

opportunities for ground passes, the ISS orbit was propagated over a month using

NOVA tracking software, as AGS4 would be in a similar orbit. Passes that never rose

20° above the horizon were not included. The propagations calculated that the AGSL

ground station would have an average ground pass time of 12 minutes, 43 seconds as

shown in the table below.

84
 Table 22: HDR and LDR Downlink Capacity
TAMU Riverside Downlink Capacity
Time Per Day per G/S 12 Min
43 Sec
High Data Rate 153600 Bps
Horizon Data Rate 72000 bps (47% of max)
Packet Efficiency (1-PLR) 80%
Effective Data Rate 57600 Bps
Pass Efficiency 50% % available passes taken
% Time Used For Payload % taken pass time used for
Data 50% payload
Per day 11.0 Mb
Low Data Rate bps
Horizon Data Rate 4500 bps (47% of max)
Packet Efficiency (1-PLR) 80%
% Time Used For Payload *% taken pass time used
Data 50% for payload
Effective Data Rate 3600 bps
Per day 1.4 Mb

To provide a conservative data rate estimate, the calculations begin by using the

data rate at the horizon, where speed of light delays are highest, and require more time

for packets and their confirmations to go back and forth. A rate of 47% was estimated

[25] for the potential horizon throughput. This data rate will improve as the spacecraft

approaches the ground station, and will maximize when directly overhead, only to begin

degrading as the spacecraft moves on towards the other horizon. Furthermore, not all

packets make it through the first time, requiring the uplink to request resending of the

missed packet. This occurs due to noise and other system imperfections. Measurements

from radio development showed that roughly 80% of packets can be reliably received

[25], and so the effective data rate is reduced by another 20%.

85
 On top of these losses, not every ground pass can be taken due to student

schedules or weather, and much of the time taken during successful passes is spent

acquiring signal lock initiating the data transfer. Data from AGS2 showed that only

about 50% of available passes were attempted or made contact, and that a little more

50% of the pass time was spent downlinking the data, reducing the total time available

for the downlink to about 25% of the total available. No pass efficiency was applied to

the low data rate system because if that became the only functioning radio, all passes

would need to be taken. The conservative margins on the calculations should ease the

burden on this high pass efficiency.

These data rates and schedule penalties are more severe than what is expected to

occur, and provide a very conservative estimate of data downlink with lots of margin.

Using these adjusted daily data rates and schedules, the following table outlines the time

required to perform the data downlink.

Table 23: AGS4 Mission Downlink Times

High Data Rate Only
Time to D/L SOH 6.6 days
Time to D/L ADCS data 1.7 days
Time to D/L VDCS 40.0 days
Time to D/L DRAGON data 12.9 days
2.04 Months

Low Data Rate Only

Time to D/L SOH 52.1 days
Time to D/L ADCS data 13.4 days
Time to D/L VDCS 314.3 days
Time to D/L DRAGON data 101.1 days
16.03 Months

86
 While the estimated time for the low data rate is long, it was within the estimated

24 months maximum lifetime for an AGS4 sized vehicle [11].

2.4.4.3. Uplink System

For uplink, and a low data rate (LDR) backup downlink, a Kenwood TH-D72A

[26] handheld transceiver was selected. As this model is a commercial product, it

required ruggedization and coating by AGSL. A particularly appealing feature of the

D72A was the internal terminal node controller (TNC), a modulator that converts digital

data signals into analog radio frequency (RF) for transmission, a feature usually not

integrated into amateur radio hardware. Another desirable feature was the transmission

(Tx) output power of 5 W, meaning no amplification is required for transmissions to

reach the ground station (see link budget). Simple USB serial connections made internal

communications and control from CDH straightforward.

Figure 59: Kenwood TH-D72A

87
 One feature that could easily be overlooked was the D72A's ability to power on

when DC power is applied to the supply input. Many electronics with a digital power

button, and not a physical switch, do not power up when they are plugged in, but instead

require activation of the button. The D72A does power up when voltage is applied,

making it controllable by the power system.

As uplinking commands does not require much data, the low data rate of 1200

baud (bits per second) was sufficient for command and control. This data rate was

selected over the 9600 baud the D72A is capable of as it operates in the 144 MHz band,

where the longer wavelength signals experience less loss and arrive at the spacecraft

with a higher signal to noise ratio (SNR) than the 440 MHz band signals the unit is also

capable of operating with. A higher SNR is desirable on the command uplink to ensure

that commands are not corrupted, and can be received when the spacecraft antenna

orientation is not favorable.

2.4.4.4. Low Data Rate Downlink System

For the primary downlink and beacon radio, a second Kenwood TH-D72A was

selected for all the same reasons as the uplink radio, and it would not require any

additional software integration efforts. Additionally, it could be used as a backup

receiver for command and control. In order to operate at the 9600 baud rate, the unit

would transmit in the 440 MHz band. The 9600 baud downlink would be used as the

primary downlink until the experimental high data rate (HDR) radio could be brought

online, and continue to be used if there were problems with the HDR.

88
 This LDR downlink system would be used for the spacecraft beacon, transmitted

every six seconds. The format of the beacon is outlined below, and it closely resembles

the SOH data being recorded.

Table 24: AGS4 Beacon Format

Beacon
Bit Size Qty. Subtotal Notes
10 Character Call Sign 80 1 80
Spacecraft Name
(AggieSat4) 72 1 72
Battery Pack Voltage 16 2 32
Battery Pack Current 16 2 32
Battery Pack Temp 16 2 32
Charge Level Indication 16 2 32
Battery Charge Status 2 2 4
Solar Panel Current 16 6 96
Safety 8 2 16
CDH Temp 16 4 64
Timestamp 64 1 64
COMM Temp 16 7 112
Position (x,y,z) 32 3 96
3 readings, one each
Acceleration (x,y,z) 32 3 96 axis
3 readings, one each
Angular Velocity (x,y,z) 32 3 96 axis
3 readings, one each
Magnetometer Reading 32 3 96 axis
Wheel Current 16 3 48
Wheel Temps 16 3 48
Torque Coil Currents 16 3 48
Torque Coil Temps 16 3 48

1212 bits/sample
152 bytes/sample

89
2.4.4.5. Experimental High Data Rate Transmitter

As 9600 baud downlink can be painstakingly slow for modern data generation

rates, which could result in lost or foregone mission data, AGS4 would carry an HDR

radio, but not count on it as the primary due to lessons learned from AGS2 [21]. For this

purpose AGS4 would fly an experimental radio unit developed by an AGSL graduate

student specifically for AGS4 and microsatellite class vehicles [25].

The radio design consists of an off-the-shelf Texas Instruments CC1101 micro

RF transceiver [27] mounted to a custom designed control board. A microprocessor

controls serial communications with the CDH, packet handling, and command and

control functions for the radio portion. This HDR radio would operate at 153.6 kbaud in

the 440 MHz band.

Figure 60: HDR Radio Unit Design (photo by Graves, Dec. 2011)

In order to make the HDR design integrate with the rest of AGS4 systems, a few

design modifications were made to the control board by the author of this thesis. The

90
DB-9 serial port connector was replaced with a latching dual row Omnetics connector,

and the power lines split off into a separate two pin latching Omnetics connector. A

voltage regulator was built in to convert the 5V power supply from AGS4 down to the

3.3V that the control board requires, as EPS would require another power output board

to produce 3.3V for just this device. Finally, the crystal oscillator that sets the serial

communication rate was replaced to bring the communications rate with CDH to a

supported rate.

Since the CC1101 at the heart of the HDR is only capable of producing a weak

RF signal, it would be routed through a 4Watt SM04093-36HS amplifier from Stealth

Microwave [28]. This power level would be just sufficient to support the desired data

rate (see section 2.4.4.9).

Figure 61: SM04093-36HS Amplifier

While differential carrier phase GPS had been originally selected as the feed-

forward technology for MR-9, problems implementing that solution arose as discussed

in a subsequent section. As the data budget clearly outlines the need for a high

throughput transmitter because of the volume of data these small satellites are

91
generating, it was decided that this experimental radio system would be AGS4's feed-

forward technology.

2.4.4.6. Crosslink

MSC-10 requires each spacecraft to exchange 50 RelNav solutions with the

other. In order to exceed this requirement, both vehicles would exchange 100. In order

to perform the RelNav calculations, AGS4 and Bevo-2 would need to exchange GPS

coordinates with each other. While it would be possible to use both spacecrafts’ ground

communications radios for this purpose, both teams decided that risk could be reduced

by utilizing dedicated point-to-point radios from the same manufacturer, designed to

work together. Radio units from Digi International were selected because they are

almost plug-and-play, and would require little software integration. As a volume

constrained CubeSat, Bevo-2 would carry the Digi Xbee chip level transceiver, while

AGS4 would carry the larger and more powerful (1 W vs. 250 mW) Digi XTend [29].

Since the XTend would have a higher transmit power as well as more sensitive

receiver and antenna, the communications distance would be increased over both

spacecraft using the Xbee.

92
 Figure 62: Digi XTend

The packets that each spacecraft would exchange with each other are outlined in

the following table.

Table 25: Crosslink Data Packet

Crosslink Nav/RelNav Solution bits
X-Position 64
Y-Position 64
Z-Position 64
X-Velocity 64
Y-Velocity 64
Z-Velocity 64
Pitch Rate 64
Roll Rate 64
Yaw Rate 64
Quaternion 256 4 qty 64 bit components
# Space Vehicles in View 16 unsigned short
GPS Week 16 unsigned short
GPS Seconds of Week 64 double
Timestamp 64
Packet Header 56 7 Byte LONESTAR Packet Header
Total 1048 bits

93
2.4.4.7. Antennas

Each of the four radios required a dedicated antenna to avoid the need for

switches or splitters, which would reduce the signal strength and introduce potential

points of failure. The design considerations for antennas were to have sufficient signal

gain in a form factor that fits within the dimensions of the airlock.

Flat patch antennas were considered for their high gain RF patterns and nearly

surface flush mounting. Unfortunately, their high gain comes at the price of narrow

beamwidth, requiring a measure of pointing control to effectively use them. While this

may be appropriate in some applications, the AGS4 team wanted communications to be

as close to omnidirectional as possible due to lessons learned from communications

trouble with AGS2. A successful mission would require that communications be able to

proceed even if the spacecraft is unable to actively orient itself. For this reason

monopole antennas were selected.

With the monopole design selected, the task became finding antennas with

sufficient gain that were short enough to clear the airlock. The antennas that came with

the Kenwood radios were too long for the airlock and offered lower signal gain than

others on the market.

Antennas that are closer in length to the wavelength of the signal usually have

higher gain values [11], with the 144 MHz signal having ~2m wavelength and the 440

MHz signal at ~70cm. In order to effectively lengthen a shorter antenna, manufacturers

will sometimes coil a conductor wire around the shaft. SRH815S dual 144/440 MHz

band antennas from Diamond Antenna were selected for both the LDR and the HDR

94
radios because of their 3 dB gain in a 6 in. length package. The 3 dB gain was sufficient

to close the communications loop with all systems (see link budget). The A09-HSM-7

antenna (that came with the crosslink radio) was used as it met clearance requirements

and was adequate for the experiment range.

A monopole antenna experiences its highest level of signal gain in the plain

orthogonal to the pole. The crosslink antenna was installed on the top of AGS4 in order

to have the highest signal strength in the same plane that Bevo-2 would be released. The

HDR and LDR antennas were installed on the rear face so that the torque coil

stabilization would orient their gain lobes towards the ground. The figure below shows

the gain pattern for the crosslink antenna. Its highest gain lobe is in the direction of

Bevo-2's release trajectory (-180° in this graph).

Figure 63: Crosslink Antenna Pattern

Similarly, the HDR antennas pattern is shown in the figure below. The gain is

consistent in almost every direction in the plane orthogonal to the antenna, making the

95
orientation appropriate for ground communications. The patterns for the LDR antennas

on the same face are almost identical.

Figure 64: HDR Antenna Pattern

All antennas were installed by ISS astronauts by hand as the SMA style

connectors are simple threaded connectors that only require to be finger tight to stay

secure.

96
 Figure 65: Antennas Installed on AGS4 (Photo by NASA)

2.4.4.8. Iridium Satellite Network Concept

A concept that was examined and considered for implementation was to leverage

the Iridium satellite telephone communications system for data return. As Iridium

has global coverage, it would be possible to retrieve data from almost any point in orbit

rather than waiting for intermittent ground passes, in a similar fashion to the NASA

TDRSS satellites. Unfortunately, as development advanced it became apparent that the

team did not have all of the skill sets necessary to integrate the satellite modem software

into the flight software, and the concept was left to future consideration. Should future

AGSL teams have the required skills, implementing satellite telephone communications

could prove to be a practical and potent communications method.

97
2.4.4.9. Link Budgets

The link budget is an analysis to determine the strength with which a signal will

arrive at the receiver, and how well it will be received with the equipment being used. If

the signal strength is greater than any noise sources, the signal-to-noise ratio (Eb/No) is

positive, and communications should be possible. A Eb/No of 3 dB or greater is a good

indicator of a strong communications link [11]. However, too high a ratio leads to the

communications becoming a source of noise to other users, so excessive power

transmissions should be avoided.

The following link calculations were performed while the spacecraft is 10° above

the horizon, which is the farthest away it will be during the ground pass. The signal will

get stronger as the separation distance closes. This analysis distance provides a

conservative estimate of the signal strength. The elevation of 10° was selected as the

benchmark for communications based on rough angular measurements of trees and

buildings that obscure the horizon.

A complete review of the elements and equations that make up these link

analyses is available in Appendix A. Additionally, the calculations for line number 17 in

each, the noise temperature, are available there as well.

98
2.4.4.9.1. VHF (LDR) Uplink

Table 26: Uplink Link Budget

LDR VHF Uplink
Transmit Power 5 W NOTES
1 37.0 dBm
2 Transmit System Losses 2 dB
3 System Amplification 0 dB
4 Antenna Gain 3 dB
5 Pointing Loss 2.37 dB 20 deg
6 VSWR Loss (VSWR of 2) 0.5 dB
7 EIRP (1-2+3+4-5-6) 35.1 dBm
Transmission Frequency, f 140000000 Hz 2.141375 m
Link Range, R (10 deg, 400
km) 1439650.4 m
Propagation Factor, n 1.0
8 Free Space Loss, Ls 138.5 dB
9 Atmospheric Absorption Lpa 0.2 dB
9b Precipitation Loss Lpp 1.0 dB
Total Propagation Loss
10 (8+9+9b) 139.7 dB
11 Receive Antenna Gain 12.3 dB
12 Pointing Loss 0.44 dB 4 deg
12b Polarization Loss 3 dB Circular
13 VSWR Loss (VSWR of 2) 0.5 dB
14 Receive System Losses 2 dB
15 Tower Cable Loss (RG-6) 2.4 dB 100 ft
Effective Received Power (7-
16 10+11-12-12b-13-14-15) -100.66 dBm
17 Noise Temperature 30.3 dB 1075.554 K
18 Boltzmann Constant -198.6 dBm/Hz/K 1.38E-23 J/K
19 Noise Bandwidth 33.8 dB Hz 1200 bps
Noise Power at 140 MHz
20 (17+18+19) -134.48 dBm
21 Eb/No (16-20) 33.82 dB
22 minimum for 10^-5 BER 13.3 dB AFSK
23 Margin 20.5 dB

99
2.4.4.9.2. UHF (LDR) Downlink

Table 27: LDR Downlink Link Budget

LDR UHF Downlink
Transmit Power 5 W NOTES
1 37.0 dBm
2 Transmit System Losses 2 dB
3 System Amplification 0 dB
4 Antenna Gain 3 dB
5 Pointing Loss 2.37 dB 20 deg
6 VSWR Loss (VSWR of 2) 0.5 dB
7 EIRP (1-2+3+4-5-6) 35.1 dBm
Transmission Frequency, f 438000000 Hz 0.684458 m
Link Range, R (10 deg, 400
km) 1439650.4 m
Propagation Factor, n 1.0
8 Free Space Loss, Ls 148.4 dB
9 Atmospheric Absorption Lpa 0.2 dB
9b Precipitation Loss Lpp 1.0 dB
Total Propagation Loss
10 (8+9+9b) 149.6 dB
11 Receive Antenna Gain 15.5 dB
12 Pointing Loss 0.44 dB 4 deg
12b Polarization Loss 3.0 dB Circular
13 VSWR Loss (VSWR of 2) 0.5 dB
14 Receive System Losses 2 dB

15 Tower Cable Loss (RG-6) 4.5 dB 100 ft

Effective Received Power (7-
16 10+11-12-12b-13-14-15) -109.47 dBm
17 Noise Temperature 25.3 dB K 341.0674 K
18 Boltzmann Constant -198.6 dBm/Hz/K 1.38E-23
19 Noise Bandwidth 42.8 dB Hz 9600 bps
Noise Power at 440 MHz
20 (17+18+19) -130.44 dBm
21 Eb/No (16-20) 21.0 dB
22 minimum for 10^-5 BER 13.3 dB FSK
23 Margin 7.7 dB
100
2.4.4.9.3. HDR Downlink

Table 28: HDR Downlink Link Budget

HDR UHF Downlink RFIC
Transmit Power 4 W NOTES
1 36.0 dBm
2 Transmit System Losses 2 dB
3 System Amplification 0 dB
4 Antenna Gain 3 dBi
5 Pointing Loss 2.37 dB 20 deg.
6 VSWR Loss (VSWR of 2) 0.5 dB
7 EIRP (1-2+3+4-5-6) 34.1 dBm
Transmission Frequency, f 440000000 Hz 0.681346 m
Link Range, R (10 deg, 400
km) 1439650.4 m
Propagation Factor, n 1.0
8 Free Space Loss, Ls 148.5 dB
9 Atmospheric Absorption Lpa 0.2 dB
9b Precipitation Loss Lpp 1.0 dB
Total Propagation Loss
10 (8+9+9b) 149.7 dB
11 Receive Antenna Gain 18.9 dB
12 Pointing Loss 0.44 dB 4 deg.
12b Polarization Loss 3 dB Circular (Max)
13 VSWR Loss (VSWR of 2) 0.5 dB
14 Receive System Losses 2 dB
15 Tower Cable Loss (RG-6) 0.1 dB 1 ft
Effective Received Power
16 (7-10+11-12-12b-13-14-15) -102.7 dBm
17 Noise Temperature 26.7 dB K 469.6618 K
18 Boltzmann Constant -198.6 dBm/Hz/K 1.38E-23 J/K
19 Noise Bandwidth 54.9 dB Hz 153600 bps
Noise Power at 440 MHz
20 (17+18+19) -117.0 dBm
21 Eb/No (16-20) 14.3 dB
22 minimum for 10^-5 BER 13.3 dB 2-FSK
23 Margin 1.0 dB

101
2.4.4.9.4. Crosslink

Table 29: Crosslink Link Budget

Crosslink
Transmit Power 1 W NOTES
1 30.0 dBm
2 Transmit System Losses 2 dB
3 System Amplification 0 dB
4 Antenna Gain 2 dB
5 Pointing Loss 0.59 dB 10 deg.
6 VSWR Loss (VSWR of 2) 0.5 dB
7 EIRP (1-2+3+4-5-6) 28.9 dBm
Transmission Frequency, f 928000000 Hz 0.323052 m
Link Range, R (100 km) 100000 m
Propagation Factor, n 1.0
8 Free Space Loss, Ls 131.8 dB
9 Atmospheric Absorption Lpa 0.0 dB
9b Precipitation Loss Lpp 0.0 dB
Total Propagation Loss
10 (8+9+9b) 131.8 dB
11 Receive Antenna Gain 2 dB
12 Pointing Loss 0.59 dB 10 deg
12b Polarization Loss 3 dB Circular
13 VSWR Loss (VSWR of 2) 0.5 dB
14 Receive System Losses 2 dB
Effective Received Power (7-
15 10+11-12-12b-13-14-15) -107.0 dBm
16 Noise Temperature 32.3 dB 1693.341 K
17 Boltzmann Constant -198.6 dBm/Hz/K 1.38E-23 J/K
18 Noise Bandwidth 42.8 dB Hz 9600 bps
Noise Power at 928 MHz
19 (17+18+19) -123.5 dBm
20 Eb/No (15-19) 16.5 dB
21 minimum for 10^-5 BER 13.3 dB FSK
22 Margin 3.2 dB

102
 While a 20o pointing error is assumed for uplink and downlink due to the coarse

control of the magnetic torque coil stabilization system, a 10o pointing error is assumed

for both spacecraft as they will both be under active reaction wheel control during the

crosslink phase, which is a more accurate attitude system with feedback control. The

XTend's 2 dB gain antenna was used in the design as the polarity was correct and it was

specifically tuned to the frequencies used.

2.4.4.10. Licensing

The United Nations International Telecommunications Union (ITU) requires that

all spacecraft be licensed to transmit by the radio communications governing authority

of the owner’s nation, in this case the Federal Communications Commission (FCC). The

ITU then facilitates coordination with all other spacecraft and nations using the same

frequency bands.

Since the HDR radio system was itself an experiment to test a new piece of

hardware, AGSL applied for and was granted an FCC experimental license (0137-EX-

ML-2012). The license call sign was WG2-XFJ, and authorized transmissions at 436.25

MHz, with a channel bandwidth of 406 kHz using frequency modulation (FM).

It was not quite as clear what radio service to file the primary LDR

communications systems under, as they were not testing new hardware designed by

AGSL. The NASA project management worked with the FCC to understand that, since

the AGS4 mission was itself an experiment, radio signals were necessary to facilitate

and the LDR and crosslink systems would also fall under an experimental license.

AGSL applied for and was granted a second license (0305-EX-PL-2014). The license

103
call sign was WH2-XGN, and authorized transmissions in three frequency bands. The

LDR uplink system was authorized to transmit at 145.98 MHz, with a channel

bandwidth of 13 kHz using FM. This would only be used for downlink if there were

problems with all other radios. The LDR downlink system was authorized to transmit at

436.25 MHz, with a channel bandwidth of 36.5 kHz using FM. Finally, the crosslink

system was authorized to transmit over the range of 902-928 MHz required for FHSS,

with a channel bandwidth of 19.2 kHz using FM.

As these two licenses only cover transmissions from the spacecraft, separate

licenses are required to operate the ground station. AGSL required that at least one

person present during a ground pass hold an Amateur Radio License in order to be the

responsible party for the terrestrial side transmissions, and that all applicable regulations

are followed.

2.4.5. Attitude Determination and Control System (ADCS)

In order to meet the requirements for 3-axis stabilization and pointing control,

AGS4 would need a system of actuators and orientation determination sensors connected

by a feedback control system. A series of trade studies were performed to determine the

suitability of hardware types for the ADCS system.

2.4.5.1. Attitude Determination Hardware

An ideal attitude determination system would consist of multiple external and

internal sensors of varying type for redundancy and to increase accuracy. However,

since resources and budgets are constrained, designers need to determine what will be

sufficient to meet the mission requirements. It was decided that AGS4 should have at

104
least two methods for attitude determination, one of which should be able to produce

results at all times.

The first type of sensor considered was a star tracker. Star trackers are cameras

that take an image of the starfield and compare it against a database of stars and their

positions in memory. When the processor has matched the image taken to the database,

adjusting for rotation, the star tracker can then determine the orientation vector of the

camera boresight and return that value to the spacecraft. The CDH then uses coordinate

transformation to gain an external orientation vector for the spacecraft. Star trackers are

often considered the best attitude sensors as they produce measurements on the level of

arcseconds, but unfortunately carry price tags close to $100k, and were outside the

budget for AGS4.

The second sensor type considered were sun sensors. Sun sensors are analog 2-

axis photovoltaic sensors that produce voltage based upon the incidence angle of

sunlight on them. Each unit is calibrated by the manufacturer and comes with a unique

voltage lookup table to translate voltage along each axis into a vector pointing at the sun.

Coordinate transformation would be used to calculate the spacecraft vector. While sun

sensors are much less accurate than star trackers, on the order of 0.5-1° as compared

with 5 arcseconds for a star tracker, they were still sufficient to satisfy the pointing

requirement at relatively low cost. For this reason, the decision was made to include one

on every face of the spacecraft. SA05 sun sensors from New Space Systems (formerly

SSBV) [30] were chosen above others for their size and mass, cost, and ability to be

delivered in time for integration.

105
 Figure 66: Sun Sensor Installed

The third type of attitude sensor investigated was Earth horizon sensors. Horizon

sensors are cameras that take images and detect where a large transition occurs, usually

from the sunlit Earth to black space, or from the dark Earth to the umbra/penumbra on

the night side of an orbit. This produces a vector to the Earth's horizon, and is useful for

calculating the nadir vector. One drawback is that horizon sensors on their own do not

produce a 3-axis solution, but work well when combined with another sensor such as star

trackers or sun sensors. The team had planned on using horizon sensors on AGS4 in

conjunction with the sun sensors until an additional handling requirement was brought

up by NASA. The astronauts and lab technicians would need handles to manipulate

AGS4 without touching other sensitive surfaces, and the only good place for them to go

106
was on the corners of the structure, right where the horizon sensors were located. Given

that the horizon sensors were second to the sun sensors, they were removed from the

design to make way for the handles, and AGS4 would proceed with only sun sensors as

the external orientation sensor.

The team also wanted there to be at least one internal orientation sensor.

Gyroscopes, while extremely accurate, are bulky, heavy, power intensive, and

expensive, none of which can be tolerated on a university level spacecraft. This left the

team examining more cost effective inertial measurement units (IMU). IMU's detect

linear and rotational accelerations, and are able to keep track of spacecraft orientation.

The accelerometers tend to wander over time, and require an update from an external

source, in this case the sun sensors, to recalibrate. The team chose a VectorNav VN-

100T IMU [31] for this role as VectorNav was founded by former students of AGSL,

and the support that they could give the program would exceed what other manufacturers

would. The IMU was capable of determining angular velocity to within 0.01°/s, and also

had a magnetic compass which provided another orientation vector, further increasing

orientation knowledge.

2.4.5.2. Attitude Control Hardware

In order to facilitate the pointing control of AGS4 to fulfill the mission

requirements, a few types of control actuators were considered.

The first was a compressed gas propulsion system, with thrusters angled to

provide rotational force. Thrust based attitude control systems have several benefits,

namely rapid angular momentum change, and an easy expansion to translational thrust.

107
However, compressed gas thrusters were quickly removed from the design space as their

drawbacks outweighed those of the other component types. Negative aspects are the

cost for reliable actuator valves, limited momentum capacity, and safety concerns from

having a pressurized vessel in the ISS crew compartment.

The second type of actuator examined was the control moment gyroscope

(CMG). CMG's operate by spinning a flywheel, often at high speed, and using gimbals

to rotate the spinning flywheel in one or two axes. The equal and opposite reaction from

the torque generated by rotating the spinning flywheel is used to point the spacecraft to

the desired orientation. CMG's are attractive because they produce a large rotational

force for a small mass, and the gimbaled axes allow a bank of CMG's to orient to any

attitude even if one or more fail. Drawbacks of CMG's are their complexity with

multiple gimbaled axes and moving parts, their high cost, and their continuously high

power consumption as they must be run constantly to avoid the stored momentum from

coming back into the bus. For these reasons, as well as a lack of appropriately sized

options available on the markets, CMG's were not chosen for the AGS4 design.

Instead, the team selected a system consisting of three reaction wheels (RW). An

RW is a simple metal flywheel spun about a single axis, without any gimbaling. While

RW's do not produce as much torque as CMG's, and are limited to producing torque in a

single axis, their reduced complexity, cost, and power consumption fit well with the

resources available to AGS4.

Three Sinclair RW-0.06 [32] units were selected to provide the pointing control,

as they were calculated to be able to effectively orient AGS4 to track Bevo-2 for the

108
RelNav experiment. Units from Honeywell were examined, but they were too heavy,

too expensive, and could not be procured in the timeframe required. Sinclair also loaned

AGSL an engineering model to practice on to reduce risk to the flight hardware. Usually

the Sinclair RW would have taken longer to procure, but AGSL happened to be offered

units originally built for another customer that cancelled their order.

The equal and opposite momentum generated from accelerating the wheel

produces a body torque about the wheel’s center. The three RW's were lined up as close

as possible with the center of gravity along each of the spacecraft primary axes to reduce

the cross-coupling effects that could induce rotation off-axis.

With both RW's and CMG's, there comes a point where the flywheel becomes

saturated with momentum, at which point this momentum must be removed in order to

be able to continue to use that actuator. Even though the net momentum for orienting

the spacecraft from one angle to another is theoretically zero because there is the

acceleration to start the slew, and an equal deceleration to stop it, friction on the wheel

bearing and minute aerodynamic forces cause losses to occur. These losses require the

flywheel to continue to spin, even when holding an orientation, and over time they

eventually require the motor to turn at full speed. At this time the momentum must be

"dumped" or "desaturated". An obvious method to perform a momentum dump is with a

propellant thruster system as is done on the ISS and other large spacecraft. However,

this type of system was rejected for the reasons stated in the above section. Instead, the

team chose to design a series of electromagnets, one per axis, that would consume power

to produce a torque against Earth's magnetic field. These devices are often referred to as

109
torque coils (TC) or torque rods, depending if they are shaped more like a loop or are

composed of a wire wrapped rod. The AGS4 team chose TC's because of the available

size of the spacecraft, and because torque produced is proportional to the area of the

loops, number of wire loops, and current run through the coils. Each TC was mounted

on one of the primary spacecraft axes, and would be able to desaturate momentum from

the other two body axes. The figure below shows the construction of the largest torque

coil.

Figure 67: Torque Coil

110
 The table below details the torque coil sizes and construction.

Table 30: Torque Coil Construction

X-Axis Y-Axis Z-Axis
Dimensions 1.77 ft x 0.66 ft 1.77 ft x 0.66 ft 1.64 ft x 1.64 ft
(0.54 m x 0.2 m) (0.54 m x 0.2 m) (0.5 m x 0.5 m)
Number of Turns 400 400 275
Wire Gauge 26 AWG 26 AWG 26 AWG
Wire Type Belden Heavy Belden Heavy Belden Heavy
Armored Armored Armored
Polythermaleze Polythermaleze Polythermaleze

Small elements of the control hardware were three 2.35" long, 0.16" diameter

metal rods, dubbed "hysteresis rods". What made these metal rods unique was that they

were made from HyMu80, a metal with malleable magnetic properties such that it is

easily magnetizable. Exposure to a magnet makes the rod a magnet, but it wears off

relatively quickly. This property acts as a sort of magnetic drag against the Earth's

magnetic field, and serves to dampen long period rotational oscillations. One rod was

included along each spacecraft axis to dampen rotation oscillations for that axis.

2.4.5.3. Attitude Feedback Control System

The orientation values from both the sun sensors and IMU, as well as the IMU

magnetometer, would be used in a Kalman filter to calculate the best accuracy attitude

estimate. A Kalman filter was chosen as it was recommended by the NASA technical

advisors to the program, and some AGSL team members had experience in

implementing them before. A Kalman filter would be useful because it uses the time

history of observations from multiple sources to achieve a more accurate solution than a

single measurement. This would be especially useful during eclipse periods where the

111
sun sensors would not produce orientation data, and the IMU alone would provide the

attitude solution. The Kalman filter would be used in conjunction with a negative

feedback control system commanding the reaction wheels.

2.4.5.4. ADCS Simulation

The ADCS system hardware, navigation filter, and feedback control system were

simulated in MatLab and Simulink to verify the system’s capability to navigate and

control the orientation. Since AGSL did not have access to a full motion feedback

control simulator, this analysis was the best way to verify the model. The simulation

concluded that the AGS4 control system would be able to orient the spacecraft for Bevo-

2 release, image the release, and track Bevo-2 for eight hours after release (battery

limited).

2.4.6. DRAGON GPS (DRAGON)

AGS4 carried a dual-channel GPS receiver designed and provided by NASA.

Initially, the design was to utilize differential carrier-phase GPS for attitude

determination. A single GPS receiver is unable to determine attitude and pointing, but

two or more can. If two receivers are spaced sufficiently apart, the orientation vector

can be calculated from the position of each antenna on the vehicle. Unfortunately, this

only works on very large vehicles like the ISS, and not on microsatellites.

Differential carrier-phase offers a solution for small vehicles. Rather than use the

position of each antenna, differential carrier-phase measures the time difference of when

a distinct GPS signal or "ping" is received at each of the antennas. Given the position of

the receiver and GPS satellite, it is possible to calculate the angle between the antenna

112
vector and the GPS satellite, providing an attitude solution. This process requires a

minimum antenna separation to be able to work, as there needs to be sufficient time for

the receiver to distinguish the time difference. The minimum separation was calculated

to be roughly 24 inches, which was just enough to implement on AGS4.

Unfortunately, due to personnel changes and the receiver not behaving exactly as

expected, AGSL was unable to implement differential carrier-phase GPS attitude

determination. Instead, the GPS receiver was used for position determination and to

collect the two orbits of GPS data as per the requirements.

2.4.7. Visual Data Capture System (VDCS)

AGS4 was required to carry a camera in order to capture visual evidence of the

spacecraft separation event, as well as to verify attitude pointing capability by imaging

celestial objects such as the Earth. Additionally, images are great tools to stir student

interest in the project.

While using a commercial grade camera was first considered, it was determined

that AGSL did not have the skill-set or equipment to ruggedize components like shutters

and lenses. Fortunately, AGSL has friends at The Aerospace Corporation through a

cooperative agreement to host an Aerospace ground antenna at the AGSL facility, and

they donated a PSCAM that they designed and manufactured for the LightSail satellites

[33]. This was preferable to the alternative as it was specifically designed for

spaceflight. The only development that AGSL had to do was create an interface board to

connect the camera’s ribbon wire connection to the computer’s serial port, and modulate

the data into the proper format.

113
 Figure 68: PSCAM

The camera has a 185° fisheye lens, and can operate in two different capture

modes, with resolutions and frame rates as laid out in the table below. Each frame for

each capture mode has a thumbnail generated to help select desired images for downlink.

Table 31: PSCAM Imaging Capabilities

Full Resolution Width 1600 pixels
Height 1200 pixels
Max Frame Rate 0.4 Frames/sec
Video Video Width 640 pixels
Video Height 480 pixels
Max Frame Rate 1 Frames/sec
Thumbnail Thumb Width 120 pixels
Thumb Height 90 pixels
Max Frame Rate N/A Post-processed

114
 Figure 69: Camera Location (red)

With the camera interface board connected, the final step to prepare for

integration was to focus the lens. This was done by taking test images, rotating the lens,

and taking another test image. When the test images were focused, the lens was staked

in place to maintain the focus setting.

The following images illustrate the size difference between the three image

capture modes. The clear blue film lens protector was still on the camera at this point in

testing, and was not removed until on the Cyclops table in orbit. While this causes some

haziness, the image is still visible, and the fish-eye effect can be seen at the edges.

These images were taken after VDCS integration to the bus, prior to panel fold up, so the

imagery is of the assembly jig and table.

115
Figure 70: PSCAM Full Resolution

116
Figure 71: PSCAM Video Resolution

Figure 72: PSCAM Thumbnail

117
 3. ASSEMBLY, INTEGRATION, AND TESTING

3.1. Assembly

3.1.1. Prototyping

Prior to component and vehicle assembly, each subsystem was prototyped using

non-flight versions of hardware to verify the system design and capability. Non-flight

versions of hardware are used such that wear and tear are reduced for the flight articles,

and so that any mistakes destroy less costly hardware. This testing also provides the

opportunity to test system interactions and develop the software to ensure that all

components are working together. In support of the type of development, prototype

level components were assembled into a configuration known as a "FlatSat".

A FlatSat is a desktop representation of the majority of subsystems of the

spacecraft that will need to be working together. For AGS4, the FlatSat consisted of a

CDH, full COMM system, camera, and emulators for the ADCS system. EPS was not

included, and was instead replaced by AC power supplies. EPS to CDH integration

would be extensively tested in other phases. This setup allowed AGSL to test all

subsystem interactions and develop the interfaces such that CDH can command and

control every aspect of the spacecraft. FlatSat was used to test the ability of AGS4 to

perform the mission operations and meet the requirements. With FlatSat development

and all other systems prototyped, components could begin to be made flight ready, and

start the assembly process.

118
3.1.2. Component Assembly

For each subsystem, the components were prepared for spaceflight separately

before being integrated into the final flight configuration. This process included hand

soldering electronics boards (where applicable), staking and coating, and function testing

of each component.

3.1.3. Soldering

Each electronic circuit board developed by AGSL, and the solar panel PCB's and

cells, were hand soldered by team members lab certified to solder. Soldering techniques

and quality assurance practices are as per NASA-STD-8739.3 Soldered Electrical

Connections [19] and AGSL-3108-200 Quality Management System [34].

Soldering consists of applying heat to the two parts being joined, and flowing a

conductive material between them to electrically and mechanically bond them. Many

commercially available solders have a high tin content to reduce potential lead exposure.

Unfortunately for spaceflight, in the cold, vacuum, microgravity environment being

operated in, tin can actually begin to grow "tin whiskers", crystalline structures which

branch out and can potentially touch other surfaces causing a short [19]. For this reason,

high lead content solder is recommended for space applications, which requires

soldering to be performed in a well ventilated environment, preferably with a fume fan.

Electronic board soldering involves placing the part in the correct orientation on

the PCB, and forming the appropriate solder connection as per [19]. Smooth, filleted

joints are desired.

119
3.1.4. Staking and Conformal Coating

Staking is the application of a thick epoxy to large soldered components, with the

intent of taking the strain of the rocket launch off of the soldered connections, and

having the epoxy become the load bearing structure. Conformal coating is the process of

applying a thin epoxy resin over the surface of boards and other components to serve as

a vapor barrier between components and the environment. It also has the benefit of

preventing certain plastics from outgassing [35], or essentially dissolving in the vacuum,

and being destroyed in the process. Staking and Conformal Coating are prescribed by

AGSL-3105-200-001 Staking and Coating Procedure Rev A [36], which is an

application specific adaptation of NASA-STD-8739.1A Workmanship Standard for

Polymeric Application on Electronic Assemblies [37]

Prior to applying any polymerics, the PCB must be thoroughly cleaned. Using

100% isopropyl alcohol and an acid brush, the board is cleaned using horizontal strokes

and washed down. The board is then rotated 90° and the board cleaned again using

horizontal brush strokes. This process is repeated twice more to be sure that all solder

residue and debris are removed, and dried with a lint free wipe. The board is then baked

at least 30 minutes at 60-65°C to evaporate any solvents.

Staking is performed by mixing a 1:5 ratio of Uralane 5753A and Uralane

5753B, along with silicon dioxide powder to thicken the mixture. After thorough

mixing, the mixture is exposed to vacuum of 0.05 Torr for 5-8 minutes. The mixture is

then placed into a syringe, the gauge of which is appropriate to the size object being

staked. A sample of the mixture is to be saved for analysis if needed. Using the syringe,

120
fillets of epoxy are applied between the body of large electronic components and the

PCB surface, avoiding the soldered connections. After 24 hours the mixture will be

cured sufficiently to perform conformal coating.

Figure 73: Vacuum Jar Setup

With all of the large components adequately secured to the PCB using staking

compound, the conformal coating barrier can be applied. The process of mixing the

resin is similar, using a 2.7:15 ratio of Uralane 5750A and Uralane 5750B. No

thickening agent is used for this resin. With the components mixed, the mixture is

exposed to the same vacuum period as the staking material, and a sample set aside. An

acid brush is the used to apply the resin to the PCB, being sure to cover all components

and fill every crevice. Care should be taken to avoid creating air bubbles, and apply thin

and even layers to both sides of the board. With the conformal coat in place, the

121
assembly is then exposed to the vacuum again to remove any trapped air bubbles on the

parts. The resin will be dry to handle in 24 hours, and fully cured in 5-7 days.

Figure 74: Conformal Coat Visible on Battery Protection Circuit

3.1.5. Locktite Adhesive

For AGS2, thread locking compound such as Locktite was not allowed by

NASA. By AGS4, NASA had changed this policy and were allowing it to be used. This

epoxy was used as back-out prevention for #0 size fasteners, and as a general bonding

adhesive on AGS4.

Locktite was the simplest to prepare adhesive used on AGS4, and is fully

described in AGSL-3105-200-002 Thread Locking Compound Procedure [38]. Another

two part epoxy, the mix ratio for Locktite is 100:17 EA9394A and EA9394B by

weight. After weighing out the correct quantities, the two are thoroughly mixed until

122
uniform in color. The working time is 90 minutes, and is best applied by spreading due

to the high viscosity. A sample should be kept for future analysis if needed.

3.1.6. SMTRS

As the structural components were fabricated by a machine shop, SMTRS did not

have to assemble these pieces. However, the first task upon receipt was to verify that

they were fabricated according to the drawings, including each hole and their spacing.

This was necessary as it was discovered that the shop that the machine shop contracts to

for anodization and Alodine treatments placed the Alodine patch for the ISIPOD

support block at a 90° angle to desired.

After the part was fixed, the next step was to install the locking Heli-Coils into

every hole that required a fastener. This was done by hand by the AGSL students using

the tools in the Heli-Coil kit for each size fastener.

The other remaining assembly task for SMTRS was to assemble the solar panel

covers, and attach Velcro to the solar panel trusses. The polycarbonate sheets had

been cut using a waterjet into a sheet with triangular grid cutouts, and a rectangular

spacer piece to keep the grid off of the solar cells. These two pieces were bonded using

Locktite epoxy, and the excess wiped away. After it had cured, a sheet of Mylar

polymer film was bonded to the outside of the grid sheet using staking compound.

Originally Locktite had been specified for this purpose, but prototyping found that the

Locktite was unable to bond to the Mylar, preventing a seal from being formed and

defeating the purpose of the panel covers. Experiments with staking compound showed

that it was able to adhere to the Mylar, and so was selected as a substitute. With the
123
film in place, the last step was using Locktite to attach strips of male Velcro to the

edges of the cover. The epoxy was applied to the hook side rather than the flat back of

the material so that when cured it could latch the cover to the matching Velcro on the

solar panel. Female Velcro was applied in matching locations on the solar panel

trusses using Locktite. The adhesive was applied to the back of the material so that the

loop side was open for attachments.

Figure 75: Solar Panel Cover

124
3.1.7. CDH

As the TS-7800 is a production unit, no soldering was required. The only

customization needed was to use the user accessible jumper pins to select the data rate

for the serial ports. The flight software and operating system were loaded onto the SD

card, and inserted into the slot on the computer. The SD card was then staked in place

along with all large components on the TS-7800 and the serial expansion boards. Stake

was applied not only to the microprocessors, capacitors, and inductors, but to all of the

serial and USB connectors to reduce the load they will experience during launch. After

the staking had cured for all boards, they were conformal coated.

3.1.8. EPS

All EPS boards were hand soldered, along with the solar cell connections and

battery connections. The boards comprising the EPS stack were all relatively

straightforward to prepare. After soldering, they were staked and coated as per the

polymerics procedure.

The solar panels were a bit more interesting. Normally solar cells are adhered to

the panel using an epoxy. This can be quite messy, and cause excess to wind up on a

cell face which cause glass breakage while trying to clean. AGSL learned firsthand from

AGS2 how frustrating this process can be. In order to eliminate this source of stress,

The Aerospace Corporation suggested that AGSL look into using double sided Kapton

tape for the adhesive, which was a project they were researching. Kapton tape has a

high enough temperature range to handle the direct sunlight and the soldering iron.

125
AGSL was one of, if not the first spacecraft to utilize this method for attaching solar

cells.

After determining that this was indeed a viable solution, cardboard templates

were designed for shapes that would provide maximum adhesion surface area while not

interfering with the cells’ connection tabs. These templates were used to cut out tape for

the 234 solar cells by hand. One side of the Kapton cut-outs was peeled off, and the

tape gently placed on the back of each cell, careful to ensure than no air bubbles were

trapped.

Figure 76: Double Sided Kapton Cut-Out

Custom solar panel soldering jigs were made to connect the cells in the proper

shape and adhere them to the PCB.

126
 Figure 77: Solar Panel Jigs

The cells were placed into the jig, and the tabs from the previous cell were

soldered to the base of the next in series. Only a small amount of solder was needed on

each tab, as a blob would have created a pressure point that would likely cause cell

breakage when pressing down to secure the Kapton to the panel.

127
 Figure 78: Solar Cells in Assembly Jig

With all of the cells connected, the backs of each of the Kapton adhesives were

removed, and the PCB for each panel was gently pressed into the back of the cells. The

jig aligned the PCB such that the cells would be in the proper locations. The PCB was

then removed, and all of the cells had adhered to it. Solder was then used to connect the

cells at the end of each string to the PCB, creating a series of 18 cells. The short wire

harness for the solar panel was soldered to the PCB at the designated terminals. This

process was the same for all 13 panels of the three types.

128
 Figure 79: Assembled Solar Panel

The batteries were a bit more difficult to assemble than the solar panels. The

first steps were to assemble the box and attach the feet and vent cover. The cell dividers

were installed, and small patches of Durette felt were placed in the bottom of each

chamber for padding and insulation. A 4" 12 gauge wire was soldered to the negative

tab of each cell, and the cell wrapped in Durette felt. The cells were then inserted into

the chambers such that the soldered lead from one reaches the positive lead of the next

and soldered together in series. Kapton tape was used to insulate all exposed leads.

The figure below illustrates the process of cell insertion and connection.

129
 Figure 80: Battery Assembly

A 26 gauge wire from a micro DB-9 connector in the box lid was run to the

positive terminal of each cell. This allowed the battery protection circuit to monitor each

cell voltage and balance the pack cells. Durette felt sheets would then cover the top of

the cells for more insulation.

What made this design difficult was that originally the 12 gauge wires for the

first and last cells in the series were to snake over across the tops of the cells in the pack

and take a sharp turn out of the hole near the DB-9 connector. Unfortunately this caused

too sharp of a bend for such a thick wire which damaged the wire. While attaching the

130
lid using the screw driver, these wires were bent beyond their limit and forced into the

sharp edge of the box lid and dividers. Although anodized to prevent conductivity, the

wires were shoved so hard into the aluminum that they chipped the coating and created a

short circuit. This produced a cascade of sparks that shot out of the almost attached lid,

dangerously and rapidly depleting the battery.

A thorough investigation discovered that the wires bending too sharply was the

cause of the problem, and the design was modified so that the pack lead wires exit the

box lid directly above the cell they are attached to. AGSL students drilled new holes in

the lid and chamfered the edges to reduce their sharpness. This would eliminate the

problems with wire chafing and being bent too far. A Teflon wire sheathing material

was added to further prevent the possibility of these wires rubbing against the edge of

the battery box. This redesign proved fairly easy to assemble, and passed all tests.

Figure 81: Prototype and Flight Battery Boxes

131
3.1.9. COMM

Since only the HDR was designed and fabricated by AGSL, this was the only

COMM PCB requiring soldering. Coaxial cables to carry the signal from each of the

radios to its antenna also had soldered connections.

The HDR amplifier did not require any components soldered, but it did utilize

power terminals where the lead wires were directly soldered on. What was unique about

the amplifier compared to the other components of AGS4 was that it was thermally

bonded to its aluminum housing. Waste heat generated would be carried away from the

electronics by conducting to this aluminum "tub" in which it was mounted. This

presented a challenge in that the underside of the amplifier could not be brush conformal

coated. Fortunately, the staking and coating procedure has an option for this use case.

As the housing is a single piece of aluminum, conformal coat could just be poured onto

the tub and the amplifier covered in a process called "potting". The assembly is then

placed under vacuum to remove all air bubbles. In this manner the amplifier is

completely coated, but maintains thermal conductivity to the housing.

As the two LDR radios were commercial units, they did not require any

soldering. Their electronics boards however would require staking and coating. After

removing the plastic housing, the two electric PCB's sit mounted in a convenient

aluminum chassis. By removing the top board, but not disconnecting the ribbon wire

connecting the two, staking and coating is a fairly easy procedure. Special attention was

given to staking the ribbon wire to ensure it would not come loose. Once cured, the

boards were replaced and the fasteners for the chassis were reinstalled using Locktite.

132
 The crosslink radio did not require any soldering. After opening up the

aluminum housing, the electronics just slide out from rails built into the housing. The

electronics were removed, and the unit staked and coated. Upon reassembly, Locktite

was used when reinstalling the fasteners.

3.1.10. ADCS

As much of the ADCS hardware is built for spaceflight, the reaction wheels and

sun sensors required no additional steps taken. The VectorNav control board and sun

sensor interface board were both hand soldered, staked, and conformal coated.

The torque coils were assembled by building a wooden jig with pegs marking the

dimensions of the particular coil shape. The jig was mounted to a wire winding device

that counts rotations of an object mounted to it. This made it simple to keep track of

how many loops of wire had been wound so far to get the desired number. Reference

Figure 67 for the build jig. Once the correct number of wire loops had been wound for

the particular coil as per Table 30, the wires were bound together into shape by tying

loops of lacing tape around them. Self tightening knots were selected to prevent the

coils from coming undone, and the knots were staked to provide an extra layer of safety.

The leads of the torque coil were then soldered to the wire harness using the Linemans

Splice [19]

3.1.11. DRAGON

The DRAGON GPS was built for spaceflight, and did not require any additional

steps.

133
3.1.12. VDCS

As the PSCAM was designed for spaceflight, the only steps necessary to prepare

the unit were to focus the lens by taking sample images and adjusting the focus. Once

the focus was found, the lens was staked in place.

The camera interface board was designed by AGSL, and so it was soldered,

staked, and coated, paying special attention to stake the ribbon wire that connected the

PSCAM to the interface board.

3.2. Integration

With all of the subsystem components prepared for spaceflight, the first step to

ensure that they survived the assembly process and are ready for integration is a function

test. Function tests were developed for each component to test the various aspects of the

hardware and ensure that they can communicate with the computer. For instance the

COMM system components would undergo a transmit and receive test, and the other

systems had similar tests that proved their functionality. Only once a component had

passed the function test could it be integrated into the spacecraft.

Integration procedures were developed for each subsystem, divided into two

phases. Phase I consisted of components that had to be installed before folding the five

lower panels of the structure together, and Phase II covered components on the +Z face

that would be the last face to be folded into place. For Phase I an assembly jig was

designed and fabricated to elevate the panels off of the assembly table, and hold the five

lower panels in their relative positions. Holes were cut in the panels so that exterior

components and some wire harnesses could be installed while attached to the jig.

134
 Figure 82: -Z Panel Assembly Jig

Panels were attached to the jig with fasteners to prevent them from slipping off. After

attaching the panels to the jig, the Cyclops EAF and spacer block with switches were

installed on the underside of the -Z panel as shown in the figure below, followed by the

two I-beam supports in the middle (see Figure 84 below).

135
 Figure 83: Cyclops EAF and Inhibit Switches

Within the Phase I & II framework, the first subsystem to be installed was CDH,

followed by EPS. The other subsystems could follow in any order with few exceptions.

The reason for this order was twofold, the first being that CDH and EPS would be used

to function test the component after install, and the second being that it made running the

wire harnesses much simpler. Since each component connected to CDH and EPS, their

wires were run from one terminus to the other, and gently placed in one of several

designated wire corridors. By placing the wires in the corridors one by one, it made it

much easier for the technician to keep track of which wire goes where and by what

route, rather than having a whole bundle of often identical looking harnesses that had to

be sorted through. Labels were made and affixed to each end of each harness as they

136
were installed just in case they ever had to be disconnected. Harnesses were run through

the wire brackets described previously, but were not clamped down until all wires were

in place. In order to minimize the number of places that wires cross panels, all the wires

running to components on a face were routed through a single corridor. This included

the wires that would run to the +Z panel, which passed through the corridor to the +X

panel, and then onto the +Z. Care had to be taken to ensure that the bend radius of the

corridor when the panel is folded does not exceed that of the thickest wire in the bundle.

The figure below shows CDH installed and the beginning of some of the wire harness

corridors, as well as the side panels in position.

Figure 84: CDH Integration

The torque coils were the first components to be installed on their panels as

harnesses for other components would need to run over them and not be pinned down.

137
They were installed using 3D printed polycarbonate brackets. The other components

were integrated as they became available and passed their function/acceptance testing.

One by one the components were integrated until all components on the lower five

panels had been installed.

Figure 85: AGS4 Pre Fold Up

After all Phase I components were installed, the ISIPOD was installed, and the

wire bracket fasteners torqued down. At this stage an integrated system test was

performed to make sure that all components installed were functioning and

communicating, as the next steps would significantly hamper the ability to remove,

repair, and replace troubled components. With all components fully functional, the side

panels were folded up, and attached together using #8-32 fasteners.

138
 Figure 86: Illustration of +Y Panel Fold Up

Remaining exterior component wire harnesses were installed after each panel

was locked in place. The other side panels were folded up in a similar manner, and

attached with fasteners such that AGS4 resembled an open topped box. The grab

handles attach via the fasteners holding the Y axis panels to the X axis panels, and were

installed at this time as the panels were fastened together.

139
 Figure 87: Folded Up Side Panels with Handles on Assembly Jig

With the side panels attached, the next step was to modify the assembly jig to

accommodate the +Z panel. A jig would be needed as the panel would be too large and

cumbersome to easily manipulate by hand, and since an expensive reaction wheel would

be installed, the risk of dropping it was too great. First the side panel jig faces were

removed to free up access space. Then the +Z jig panel was attached to the assembly

table.

140
 Figure 88: +Z Assembly Jig Panel

This jig panel attached to the +X side panel via a piano hinge, which would allow the +Z

panel to be lowered into position slowly and in a controlled fashion.

Figure 89: Illustration of +Z Jig Attachment

141
 The +Z panel was attached to the jig plate, and the components for that face were

installed. The wire harnesses had previously been run from CDH and EPS through the

wire corridors up to the panel. This was the one case where running the wires before the

components made sense as there were few enough components that labels were not

cumbersome, and there was not sufficient room to maneuver hands inside the bus when

the panels were folded up. The wire harnesses were routed to their respective

components and attached, and the wire brackets torqued down to hold the wires.

With all components installed and connected, a final full system function test was

performed to ensure that all components were not damaged during the integration

process, and that the system as designed is fully functional and capable of meeting the

mission requirements.

Prior to attaching the +Z panel, a final round of staking compound was applied to

numerous locations throughout the bus. Upper management performed this procedure

and used their discretion to determine what areas would benefit from it. Any place

where wires went through a panel were staked to prevent movement and chafing against

the edges. Places where wires had slack such as near panel crossings, and gaps around

wire brackets were filled with staking compound to prevent movement. The knots of

other locations where lacing tape had been used were also staked, and any other place

where wires could potentially rub against a hard surface.

The last step to structural integration was the attachment of the +Z panel. To

facilitate this, pins locking the jig structural members to the table were removed, and the

entire jig plate hinged over following the piano hinge track. The hinge had been drilled

142
and cut so that the top panel lined up with the rest of the structure, paying careful

attention to the wire crossing the panels.

Figure 90: Illustration of +Z Panel Fold Up

The fasteners holding the panel to the jig were removed, and the jig folded away.

The panel was then attached to the rest of the structure using 44 #8-32 fasteners, and the

piano hinge and jig were detached. All of the structural components were now

integrated, and another component function test performed to verify all are functional

after the procedure.

The next step was to prepare the solar panels and their truss backing structures.

A thin strip of Locktite was applied to the surfaces of the truss structure that would

contact the solar panel PCB in order to bond the two and prevent the PCB from

oscillating independently during launch. Bolts and nuts were placed through the

mounting holes to properly align the pieces. Once cured, the entire assembly was

integrated onto the spacecraft bus in the proper positions, and attached with fasteners.

Washers were used to distribute the force of the socket head cap screw over a larger area
143
of the PCB to reduce stress and help prevent fractures. In one bolt location the solar

panel assembly process had shifted the cells approximately 1/32" toward the bolt hole

and created an interference between the corner of the cell and the washer. Since the

washer was deemed to be a fairly important stress reducer, a Dremmel® tool was used to

cut a piece out of the washer such that there was clearance for the cell, and surface area

to distribute the fasteners pressure. The wire harnesses were attached to their mating

leads and fastened to the spacecraft bus. Staking compound was applied to keep the

wire harnesses in place along the positions where they leave the solar panel. With the 10

upper and side face solar panels installed, the covers that had been made were attached

using the embedded Velcro strips.

Figure 91: Bus with Solar Panels and Covers Installed

144
 Since the -Z panel had been unavailable for access as it was structurally

supporting the bus, it needed to be flipped over so the bottom panels could be integrated.

This would be done using the foam packing clamshell pieces as they provided sufficient

protection and supporting area to hold the vehicle without damage. It would also serve

as the first fit check for the foam with the solar panels and covers, though a preliminary

fit check had been performed once before as shown below.

Figure 92: Foam Fit Check

Using four of the stronger AGSL members, a maneuver of lifting AGS4 by the

attached handles, rotating it 180° upside down, and gently lowering it into the foam was

executed. The biggest problem experienced during this maneuver was that the handles

were a bit small to accommodate two peoples’ hands during the transition, sometimes

145
making it difficult to transfer leadership of a particular handle. With AGS4 transferred

to the foam packing, the bottom face solar panels were applied.

Figure 93: -Z Solar Panels Installed

From this point on, the assembly jig could no longer support AGS4, so it was

dismantled to clear the assembly surface. Now, any time that AGS4 was to be removed

from the packaging it would reside on the test stand, after being transitioned there by

reversing the four person lift and rotate maneuver. The test stand was created by

repurposing the prototype microsatellite bus from a previous project, as the triangular

grid design gave the aluminum a good amount of strength, reinforced with pipe to

prevent sagging. This structure was tested to hold a heavier object than AGS4, and so

was deemed safe for use. The structure had an opening at the top that was large enough

to allow the Cyclops EAF to pass through, but narrow enough that the Cyclops adapter

spacer would rest on the surface supporting the vehicles weight.

146
 Figure 94: AGS4 Test Stand (behind AGS4)

When running wires through brackets as discussed above, if the number of wires

is not sufficient to fill all of the space, the wire bundle was wrapped with Kapton tape

to fill the gap to reduce vibrations. In the case of the steel pipe brackets, the 90° metal

edges could easily cut through the Kapton tape and wire insulation, risking an electric

short. To prevent this from happening, wires running through these brackets were

wrapped in several layers of 3M® glass cloth tape, which is very abrasion resistant and

would protect the wires from chafing. Since the polycarbonate 3D printed brackets had

smooth edges made of plastic, this abrasion tape was not necessary. Future projects

should consider using 3D printed brackets for reasons such as these, and if there is a

printer in-house brackets can be tailored to each bundle without the need for filler wrap.

147
 All load bearing fasteners used on AGS4 were required to be certified of a grade

of sufficient strength and temperature durability to meet the loads expected. This

certification and chain of custody process increases the price of fasteners orders of

magnitude above fasteners from the hardware store. Counterfeit fasteners are a

surprisingly big business, and several programs have had to reassemble spacecraft after

purchasing them. National Aerospace Standard (NAS) fasteners were chosen from a

reputable supplier for all load bearing fasteners because of their traceability and

certification process. NAS fasteners follow the U.S. customary system of measurement

and fastener specification, #X-YY, where the X represents the diameter of the fastener in

64th of an inch, and Y indicates the number of threads per inch. Hardware store grade

non load bearing fasteners were used in some locations, including the metric system

fasteners. This was not a problem though as representative samples of every type of

fastener used would be destructively tested by NASA to ensure they meet their minimum

strength requirement, were made of the appropriate material, and did not have an

inappropriate surface treatment.

Each fastener installed into the bus or a component was torqued to a specific

value calculated for that application. Under tightening can cause components to vibrate

excessively, and over tightening can cause damage to components, strip the threads, or

place the system under enough spring tension that causes such as thermal expansion can

cause a violent fracture. Unfortunately, calculating the proper torque per fastener is not

a straightforward task. Factors such as the mass of the objects being fastened, the

temperature extremes the joint will experience, the size and number of fasteners, fastener

148
length, material types, and whether any lubricants are used are a few examples of the

parameters that must be accounted for to calculate the fastener torque. Fortunately,

NASA developed a standards manual [39] on how to calculate the appropriate fastener

torque. The details of how to perform these calculations are captured in Appendix C.

The correct torque was calculated for each fastener, and applied during the assembly

process.

3.3. Integrated Testing

With AGS4 fully integrated, the process of testing the assembled spacecraft

began. Integrated system tests were performed to emulate the spacecraft functions. For

instance, EPS tested battery charge and discharge, torque coil output and magnetic field

strength, the hazard control module's ability to prevent Bevo-2's command release before

timer expiration, and output boards by providing power to all components for their own

system tests. CDH was tested in a similar manner by facilitating the tests of the other

subsystems. COMM simulated a crosslink test and data downlink pass. VDCS and

DRAGON systems were used to capture images and GPS data to verify their functioning

and controllability by CDH, and were downlinked over radio. ADCS systems verified

that CDH could command the RPM of the reaction wheels, activate the torque coils, and

get accurate readings from the Sun sensors and IMU, however, a full feedback control

test was not possible in the lab environment. With AGS4 fully built and functional, the

process of integrating Bevo-2 could begin.

A full crosslink test had been successfully performed by the AGSL and UT teams

previously using prototype hardware. Now, the UT team brought Bevo-2 to AGSL

149
where the test would be repeated using the flight articles. The test was successful, as the

process had worked out all the issues during the prototype tests, and both vehicles were

now ready for final preparations. Prior to installing Bevo-2 into the ISIPOD, its thruster

fuel tank needed to be filled with chlorofluorocarbon refrigerant. This operation was

performed by the UT lab.

With Bevo-2 fueled and ready, both spacecraft could now be integrated. A

wrench was used to actuate the manual release of the ISIPOD door, and gravity as well

as spring force held it open. Bevo-2 was loaded onto a custom designed Teflon tray

that lined up exactly with the ISIPOD. While one UT student held the tray in place,

another pushed the CubeSat into the ISIPOD opening, compressing the deployment

spring. The spring was strong enough that it could eject Bevo-2 in microgravity, but on

Earth there was significant resistance making it easier to hold in place. With Bevo-2

sitting flush with the ISIPOD opening, the door was closed and forced shut. The

latching mechanism locked the door into place, and Bevo-2 was integrated. ISIPOD

contained zero separation force connectors on the inside that connected Bevo-2 to a DB-

9 and DB-15 port on the outside of the deployer structure.

150
 Figure 95: ISIPOD Payload Access Connectors

AGSL had installed a wire harness that brought these pins out to a second computer

access port on the side of the AGS4 bus. This test port was used to connect with Bevo-2

and verify that the computer system could be communicated with after install, and that

the batteries could be charged externally or by AGS4. All other systems of Bevo-2 were

three fault tolerant inhibited for safety just like AGS4. The test ports on the spacecraft

are shown in the figure below. The ISIPOD door was the held shut using a locking cable

tie to mechanically ensure Bevo-2 could not be released, and the tie would be cut off on

151
-orbit prior to release. With the two spacecraft integrated, the unit could now undergo

final NASA acceptance testing.

Figure 96: Computer Access Ports for AGS4 (Circular) and Bevo-2 (DB-9/DB-15)

With the LONESTAR spacecraft fully assembled and integrated, human factors

specialists came to AGSL to verify that the post-launch preparations could be performed

by the astronauts on orbit without injury or damage to the vehicle or station. In case of a

failure of the ISS deployment equipment, AGS4 would have to be retrieved by

spacewalk, so making sure that nothing would damage a space suit was critical. This

process was also used to verify and refine the astronaut handling procedure and

preparation. Keep out zones such as the solar panel surfaces and ISIPOD release

mechanism were identified. First, the specialists followed the preparation procedure to

install the antennas, and remove the solar panel covers. They then wiped down the

corners, edges, and all other surfaces with a soft cloth looking for any places that it

152
might snag, potentially cutting an astronaut or space suit. They also verified that the

covers were sufficiently transparent to identify any cell breaks.

Figure 97: Human Factors Testing

The handles were checked to verify that a space suit gloved hand could

adequately use the maneuvering handles, with or without the solar panel covers in case

any had to be left on for deployment. They also tested the ability to reattach the covers

in case there was the need to abort deployment and re-stow AGS4 in the foam packaging

and M01 cargo bag. After passing all of these tests, the unit could proceed to vibration

testing to implement the simulations performed in section 2.4.1.5.

In order to verify workmanship and structural integrity, the integrated AGS4 and

Bevo-2 would be vibrated along each axis to the loads and duration outlined in section

2.4.1.5. The spacecraft, packaged in the foam clamshell and wrapped in the M01 cargo
153
bag, was strapped down to the shaker table using cargo straps. Since the table only

moves in one axis, the test was repeated three times, rotating which face was in contact

with the table. The figure below shows one of the tests being performed.

Figure 98: Dynamic Random Vibration Testing

Basic function tests of each component were performed on AGS4 after every

axis run, and the UT team checked Bevo-2's functionality as best as able with the limited

functionality available while in ISIPOD. For the Z-axis and X-axis runs both spacecraft

performed flawlessly. Unfortunately, after the last run on the Y-axis the HDR radio was

unresponsive, indicating a problem. As there were no loose pieces inside the bus, it was

believed to be either a wire issue or an electronics issue with its control board.
154
 Having been given a passing grade from a safety standpoint, AGSL brought the

vehicle back to AGSL to investigate. Removing the +Z panel was not possible as doing

so would break the "as tested" configuration and void the testing results. Non structural

members such as the solar panels could be removed to investigate. When investigating,

no wire damage could be found to repair, and the HDR failure deemed to be a problem

with the control board. As there was insufficient room to maneuver tools and hands

inside to remove and replace it, the decision was made to proceed with only the LDR

radios as the mission could still be completed with a longer downlink period.

AGSL continued to refine and test the flight software up until it was time to bring

AGS4 to NASA Johnson Space Center (JSC) in Houston for delivery. The software was

installed using the computer access port. AGS4 was taken to JSC in foam and an M01

bag, where it was transferred into fresh foam and a different cargo bag. This was done

as a precaution since these items had been outside of NASA care, and they needed to

ensure that their properties were as modeled. NASA took possession of the package for

transport out to Kennedy Space Center, where it was integrated into the cargo launch

vehicle.

155
 4. LAUNCH, DEPLOYMENT, AND OPERATIONS

4.1. Launch

AGS4 was launched to the International Space Station (ISS) aboard the Orbital

ATK commercial resupply vehicle OA-4 (Cygnus) inside the foam packed soft-stow

M01 bag as cargo. The rocket was the ULA Atlas 5, which experienced 3 scrubbed

launch attempts before finally being able to liftoff on the morning of December 6, 2015.

OA-4 rendezvoused with ISS three days later on December 9, where AGS4 was

unloaded with the other cargo.

4.2. Deployment

AGS4 was removed from the M01 bag and packing foam to be installed onto the

Cyclops payload deployment system on January 27, 2016. Astronauts Scott Kelley and

Tim Peake donned gloves and static electricity grounding straps to begin the AGSL

unboxing and installation procedure.

First, the upper half of the packing foam was removed and the lower solar panels

inspected for cracks and breaks during launch. If breaks had been found, the Mylar

covers would have been left in place to avoid introducing tiny glass shards into the

weightless cabin environment. After verifying that there were no cell breaks, the

Mylar coverings were peeled away, and AGS4 was removed from the lower half of the

foam packaging. During this process, one of the astronauts received a static discharge

when touching the metal structure of AGS4, due to the charge buildup from rocket

vibrations and the foam packing. It was determined that since the flow of charge was

156
from AGS4 to the astronaut rather than the other way around, that it was unlikely that

this spark caused any damage.

Figure 99: Astronauts Removing Lower Solar Panel Covers (Photo by NASA)

Figure 100: AGS4 Removed from Packaging (Photo by NASA)

157
With the bottom cells uncovered, the astronauts carefully placed the Cyclops knob into

the Cyclops table and locked it into place. When firmly seated, the four inhibit switches

in AGS4's Cyclops attachment knob are engaged and prevent activation.

Figure 101: AGS4 Install onto Cyclops Table (Photo by NASA)

With AGS4 secured to the table, the astronauts carefully inspected the other 10

solar panels for cell breaks. Again, if any had been found broken that cover would

remain in place for deployment to prevent cabin contamination. Finding no breaks, the

covers had their Velcro attachments undone and the covers removed. With the covers

removed, the four communications antennas were threaded onto their attachment points

by hand.

158
Figure 102: AGS4 Solar Panel Covers Removed and Antennas Installed (Photo by
NASA)

With the antennas in place, the last step to prepare AGS4 for deployment was to

move the inhibit key from "SAFE" to "ACTIVE". An Allen wrench screwdriver was

used to attach the two captive #2-56 screws in the inhibit to prevent it from coming loose

on-orbit and de-activating the spacecraft. With the inhibit key now in place, the only

thing preventing AGS4 from activating were the four inhibit switches in the knob, which

would close on separation. A custom timer board would then activate and begin a clock

that counts up, and closes another electric inhibit switch upon expiration allowing the

remainder of the spacecraft to power up. This timer is to prevent AGS4 from activating

in proximity to the ISS. The timer counts up instead of down so that in the event of a

power fluctuation and the clock resetting to zero, the count restarts rather than suddenly

thinking that time has elapsed.

159
 Figure 103: Inhibit "Active" (Photo by NASA)

With the inhibit key installed, the Cyclops tray with AGS4 was slid into the Japanese

Space Agency JEM Airlock, where it would remain for the next two days. The delay

was to accommodate the ISS crew schedule.

Figure 104: AGS4 Insertion into JEM Airlock (Photo by NASA)

160
 Figure 105: JEM Airlock (Photo by NASA)

When two days had elapsed (January 29, 2016), the JEM was depressurized, and AGS4

moved outside into an area called the "porch". Over the next several hours, the Special

Purpose Dexterous Manipulator "Dextre" maneuvered into position, attached to the

Cyclops table, and moved AGS4 into deploying position.

161
Figure 106: AGS4 and Cyclops Table Moved onto Porch (Photo by NASA)

Figure 107: Dextre Moving AGS4 to Deploy (Photo by NASA)

162
Figure 108: AGS4 Pre-Deploy (Photo by NASA)

163
 NASA determined that the proper release angle for payloads is 45° down (radial)

and 45° to the side (cross-track) off of the ISS velocity vector to ensure that released

objects will not come back and possibly conjunct ISS on the next orbit. Once Dextre

had AGS4 in this orientation, the Cyclops table was activated, and the pre-loaded spring

force shoved AGS4 away from Cyclops.

Figure 109: AGS4 Release (Photo by NASA)

164
 Figure 110: AGS4 Free Flying (Photo by NASA)

4.3. Operations

After release from Station, NORAD began tracking AGS4 within a few days, and

Dutch Amateur Radio operators were the first to track and detect the beacons.

Unfortunately, the AGSL team was having trouble compensating for Doppler shift at

low elevations, and was unable to establish communications in a timely manner. It was

calculated that tuning the receiver down by 15 kHz while the spacecraft was approaching

(blueshift), and up by an equal amount when departing (redshift) during a ground pass,

or also widening the receiver channel, that the Doppler effect could be overcome. This

proved easier in theory than in practice with the manually controlled ground station radio

being used.
165
 NORAD had begun tracking two distinct space objects, with the only likely

answer being that Bevo-2 had been prematurely and inadvertently released. A fault tree

analysis was begun to determine the cause of this unintended release, focusing on the

release control electronics, and the structural integrity of the ISIPOD. The analysis came

up with six options that could have caused the release.

Option 1: Release command prematurely sent. This was deemed unlikely

because the timers on the hazard control module create an open circuit condition for

power to be applied to the release mechanism, and the required 31 days had not elapsed.

Option 2: Timer values changed due to bit flips. The time value that had to

elapse before the hazard module would close the circuit, 2,678,400 s, was stored in four

separate memory locations to prevent the value from being changed. Testing with the

hardware showed that even with one memory value changed, the highest time value

must still be reached before the transistors can close, and so this scenario was deemed

unlikely.

Option 3: All four timer memory locations were changed to lower values by high

energy particles. In order for this to occur, the spacecraft would have to be exposed to a

significant amount of radiation, and it is estimated that this amount would have also

damaged the transistors, causing them to become stuck in the off position. This situation

was deemed unlikely because this would have caused the system to fail safe, and Bevo-2

would never have been able to be released.

Option 4: Sneak short circuit to ground. This was deemed unlikely as ISIPOD

was inhibited on the ground leg by the hazard control module, and by EPS on the

166
positive side. Even if there was a sneak ground path, power would have still had to have

been applied by EPS . As this system was designed to be three fault tolerant, this was

deemed unlikely.

Option 5: Thermal expansion and contraction caused the ISIPOD door

mechanism to fail. This was deemed unlikely as the ISIPOD design was thermal

vacuum tested for six cycles at -40-+90 °C, and flight articles are tested to two cycles.

This acceptance testing showed that the design will not fail due to expansion and

contraction.

Option 6: Repeated thermal cycles caused stress on the ISIPOD latching

mechanism, causing it to fail. This scenario was deemed the most likely cause of failure

as the roughly 34-38 thermal cycles experienced before the release were well in excess

of what it was tested for. ISIPOD type dispensers are usually mounted to the outside of

the launch vehicle, and only experience a handful of thermal cycles before activation.

Additionally, the UT team cycled the mechanism several dozen times for analysis,

testing, and video demonstrations, which far exceeds what AGSL policy dictates could

be done with flight hardware. These factors combined make release mechanism due to

thermal stress the most likely cause of the unintended release.

Unfortunately, the trouble for AGS4 did not stop there. After six days of

receiving and tracking the beacons, AGS4 went silent. Re-boot commands, and attempts

to switch radios were unsuccessful. Data from some of the beacons indicated that there

may have been some data corruption issues due to radiation effects. The FlatSat was

used to duplicate the problem, and it was found out that AGS4 had a critical software

167
flaw. After about 52 hours the memory would fill up, forcing a re-boot. EPS would also

power cycle CDH if it did not receive signals after an hour. This memory filling

problem, combined with the potential radiation corruption of memory, put CDH into an

unrecoverable state. This problem could have been caught in ground testing, but tests

had never been run for more than 24 hours. Future missions should run the whole

mission end-to-end on the ground before launch, not piecemeal and stitching segments

together.

After it was deemed unrecoverable, AGS4 operations ceased, and the process of

identifying all of these potential faults and failures began to help the lab learn from past

mistakes and apply them to future programs.

168
 5. CONCLUSIONS

Even though AGS4 was unable to complete the Relative Navigation experiment,

and subsequently experienced total loss, the students of AGSL learned more from the

design and assembly processes than in much of their college classes. Working under

NASA supervision to produce a product that had to stand up to all the real-world

requirements and constraints gave AGSL students a leg up when competing for

employment, and helped to produce more qualified engineers with design and hardware

experience beyond the theoretical to improve the capability of the U.S. aerospace

industry.

5.1. Requirements Design Evaluation

The students designed AGS4 to meet the mission requirements derived from the

mission objectives set forth by NASA. While unable to implement most of them on-

orbit, they are the design criteria that shaped the design of AGS4, and the following

summarizes how the spacecraft design criteria would satisfy these mission requirements.

MR-1 was satisfied by the modeled dimensions of AGS4, including antennas and

solar panel covers, fit within the NASA provided CAD envelope of the JAXA JEM

airlock, and was verified by the deployment through that airlock.

MR-2 was satisfied by the design mass of both spacecraft being under their

allowances, resulting in a combined configuration of <100 kg. This value was verified

by NASA at acceptance.

169
 MR-3 was satisfied by AGS4 coming in under the allowance of 50 kg including

solar panel covers, excluding Bevo-2. This was verified by weighing AGS4 prior to

Bevo-2 integration.

MR-4 was satisfied by the Bevo-2 spacecraft weighing significantly less than the

allowed 35 kg, verified by weighing the spacecraft prior to integration with AGS4.

MR-5 was satisfied by both labs developing a program schedule, and adhering to

it with alterations, delays, and changes as dictated or approved by NASA management.

MR-6 was satisfied by student management from both labs having a weekly

teleconference with NASA management to provide updates and make requests of their

partner or management.

MR-7 was satisfied by the AGS4 design utilizing reaction wheels and magnetic

torque coils to stabilize and actively orient the spacecraft.

MR-8 was satisfied by both labs developing and establishing a format and

protocol for crosslinking between the two spacecraft, verified by the crosslink test and

demonstration.

MR-9 was satisfied by the inclusion of a high speed radio modem developed by

AGSL, which would be necessary to downlink the large volumes of data generated by

ARD maneuvers in future missions.

MR-10 was satisfied by AGS4 utilizing the RelNav solutions exchanged between

both spacecraft to actively track Bevo-2 as it separates from AGS4, testing algorithms

and capabilities to orient a spacecraft for rendezvous and docking in subsequent

missions.

170
 MR-11 was satisfied by the inclusion of multiple radios in the communications

system, two capable of bi-directional communications, and one high speed downlink

only radio. This system was verified by testing prior to launch.

MR-12 was satisfied by the inclusion of a dedicated crosslink radio on both

spacecraft for communications with the partner, verified by the crosslink demonstration

tests.

MR-13 was satisfied by developing a protocol and format for exchanging GPS

position states with the partner spacecraft using the crosslink radio, verified by the

crosslink demonstration tests.

MR-14 was satisfied by AGS4 carrying the PSCAM and developing the

sequence to image Bevo-2 as it was deployed. Space in the data budget was dedicated to

downlink this data.

MR-15 was satisfied by the inclusion of the DRAGON GPS receiver in the

design, and its use as the means of position determination for RelNav calculations, and

in the active tracking of Bevo-2 to test GPS usefulness for ARD.

MR-16 was satisfied by the mission concept of operations dedicating a portion of

the mission to capture and downlink two orbits of DRAGON GPS data.

MR-17 was satisfied by the inclusion of the ISIPOD mechanism built into AGS4,

and verified by the testing of AGS4's ability to actuate the release mechanism.

MR-18 was satisfied by the inclusion of the PSCAM in the design of AGS4, and

the mission concept of operations including the capturing of the Bevo-2 release, as well

as tracking and imaging the Earth and Moon.

171
 MR-19 was satisfied by development of the RelNav protocol and exchanging of

GPS solutions via crosslink radio, and verified by the crosslink demonstration tests.

MR-20 was satisfied by the development of the data plan including the stored

spacecraft health data to be downlinked and beaconed from AGS4.

As demonstrated above and in this document, AGS4 and the mission it would fly

were designed to satisfy these 20 mission requirements.

5.2. Mission Summary

To meet the mission requirements, AGS4 would perform a series of actions.

After release from the ISS, AGS4 would collect orbits of raw and carrier phase GPS

data, and downlink as able. After charging Bevo-2, it would be released, and the two

spacecraft would exchange navigation solutions with the other. The release would be

captured on video. Additionally, AGS4 would track Bevo-2 using the exchanged

navigation solutions to demonstrate attitude control for future ARD missions. The

remainder of the spacecraft life would be spent downlinking the data and images from

the experiment.

5.3. Spacecraft Design Summary

To meet the mission requirements, the final design was based around a

24x24x11" box shaped bus containing a CubeSat deployment mechanism. The complete

power system was made in-house, to include battery packs, solar panels, regulation, and

distribution systems. Command and data handling was performed by a single board

computer with supplemental serial expansion boards. Communication with the ground

was facilitated by two bi-directional commercial handheld radios and a custom made

172
high speed downlink radio, while communications with Bevo-2 were handled with a

commercial point-to-point radio. Spacecraft attitude control and stabilization was

facilitated through three single axis reaction wheels and magnetic torque coils. Attitude

determination was handled through a system of six sun vector sensors, and a three axis

linear and rotational accelerometer/magnetometer IMU. Position navigation solutions

were generated from the NASA provided DRAGON GPS receiver. Finally, visual

imagery for the mission would be captured through the PSCAM provided by the

Aerospace Corporation.

5.4. Assembly Summary

AGS4 was assembled on a panel by panel basis after the components for that

panel had been assembled and prepared for spaceflight. After the -Z panel, and X and Y

panels had their components installed, the panels were folded together and fastened such

that it made an open top box. The +Z panel was then assembled, and folded into

position and attached once complete. Then, the solar panels and their covers were

installed on the outside of the spacecraft. Testing was then performed to ensure that the

spacecraft was functional and spaceworthy, including the random dynamic vibration

testing. AGS4 was then transferred to foam packaging and delivered to NASA for

launch vehicle integration.

5.5. Lessons Learned and Recommendations

This section is a list of lessons learned that others developing their own space

mission or designing their own satellite might find useful. The topics are presented in no

particular order.

173
1. As much as possible, have external interfaces identified and fully defined

before designing hardware. The Cyclops table and EAF were developed

in parallel with AGS4, which caused delays and necessitated design

changes and modifications.

2. If using commercial grade computer components, at the very least use

radiation hardened or tolerant data storage.

3. During integration and testing, unshielded power wires were producing

electromagnetic interference on data wires, preventing communications.

This was remedied by twisting the power wire conductors into a helix so

the ground wire provided some shielding. If possible, use shielded wires

for power and data, or run them in separate corridors. If unable to do

either, at least twist the wires of each harness together to provide some

shielding.

4. Always be cognizant of the bend radius of a wire bundle, as exceeding

this can cause wire damage.

5. 3D printed wire brackets are an inexpensive and easy way to secure wires

and cables. If the design lab can obtain their own printer capable of

printing a suitable material like polycarbonate, then brackets can be sized

to the particular location, eliminating the need to wrap the wires with tape

and such.

6. Perform full end-to-end mission script testing, rather than testing each

segment separately, and saying that the flight software will implement

174
 each segment. Doing this would have exposed the memory allocation

flaw in AGS4, which could have been addressed.

7. Take high quality pictures of every component at every stage in

development, not just final pictures, label them, and categorize them

appropriately. This can help to track down when problems occurred.

8. Adhering solar cells with double sided Kapton tape is a significant time

saver compared to traditional epoxies, and is far less likely to break cells.

Several of the 24 cells for AGS2 were broken during fabrication and had

to be replaced, while only one of the 234 cells on AGS4 broke during

fabrication.

9. Create an easy to track system for the number of times a soldered

connection has been re-worked. AGS2 and AGS4 simply wrote it in the

certification log, which did not make it easy to keep track of.

10. Using a disposable polypropylene cup instead of a glass or polypropylene

beaker eliminates the need to clean and wash the beaker after mixing

epoxies, as it can simply be thrown away when done.

11. The experimental radio license is probably the simplest to apply for and

be granted, and is a good avenue to investigate for communications

licensing.

12. Make software and people capable of writing it an early and integral part

of the design process. The AGS4 team was almost entirely aerospace

engineering students, and as such there was a distinct lack of software

175
 knowledge, especially around embedded hardware systems. Having more

software focused team members with hardware experience could likely

have prevented the satellite from becoming unresponsive.

13. If re-using the HDR radio design, redesign the control board such that the

CC1101 is either integrally mounted to the control board, or mechanically

fastened. On AGS4 it was attached using the friction of the pin

connectors and epoxied, which is believed to have been the most likely

reason that the HDR did not survive vibration testing as all other

electronics had no problems.

5.6. Personal Contributions

The author worked at AGSL from Fall 2008 through Summer 2015, and had the

opportunity to work on both AGS2 and AGS4. For the AGS4 program, he filled the

roles of communications subsystem Team Lead, Chief Engineer, and Graduate Lab

Manager. This thesis outlines the decision making process and system design for all

aspects of the spacecraft, where the more detailed sections represent areas where the

author had significant input. These areas, along with other significant roles,

responsibilities, and contributions are outlined in the list below:

1. Oversee subsystem and vehicle design, integration, and testing.

a. Oversee and verify thermal and structural analysis.

b. Verify EPS system designs.

c. Lead redesign and retrofit efforts for battery boxes after the short

circuit accident.

176
 d. Design the solar panel covers.

e. Develop the method used to attach solar cells to the PCB panel

using Kapton tape.

f. Design the concept of using 3D printed wire brackets.

g. Develop communications architecture and apply for FCC

licensing.

h. Modify the HDR design to interface with AGS4.

i. Review and verify schematics and mechanical drawings.

j. Develop the fastener torque calculation procedure and calculator.

k. Perform and oversee integration of components into the vehicle

assembly, and the performance of integrated testing procedures..

2. Oversee Lab quality assurance and requirements verification

a. Train students in soldering techniques, polymerics, and other lab

procedures.

b. Oversee Lab safety practices.

3. Interface with NASA management to execute tasks and

recommendations.

a. Responsible to ensure AGS4 compliance with NASA safety

requirements.

b. Interface with UT team on crosslink and other interfaces.

177
5.7. Final Evaluation

Participating in the design and fabrication of AGS4 provided students the unique

opportunity to work on an actual spacecraft in a real world engineering environment

with NASA. Students learned how to distill mission requirements into a spacecraft

design and concept of operations to satisfy those requirements. They also learned how

to perform trade studies to narrow design choices to those that best satisfy the

requirements, and how to perform engineering analysis on real-life systems which is the

next leap forward from the theoretical analyses taught in engineering classes. Students

learned how to fabricate components and systems according to designs and

specifications, learning how to perform the roles of the technician. And very

importantly, students were exposed to the regulatory environment involved when

working with NASA or other Government entities, how to adhere to necessary design

standards, and that such standards exist for almost every aspect of spacecraft design.

Students from AGSL are more prepared to go into the workforce and hit the ground

running than their counterparts, as they have real world engineering and space mission

experience under their belt. Such prepared workers are vital to the United States

continued leadership in space, and ensuring freedom of action in space for all.

My time at AGSL remains some of the best and most influential of my life,

where I was able to learn by doing, and sometimes failing, which shaped me into the

engineer I am today. Where else does one have the opportunity to participate in

designing, building, and flying two spacecraft missions while at college? Being allowed

to make mistakes and learning from them teaches one far more than any textbook ever

178
could. AGSL is a true gem of the Texas A&M engineering program, and every student

across the country should be so lucky to get to participate in such a program.

Figure 111: Signatures and Aggie Rings of AGSL Members and NASA
Management displayed on AGS4. Also shown (top) is the Aggie Ring of Civil
Engineering graduate Patrick Brand (’81) whose father, Vance D. Brand, carried it
into space as commander of Space Shuttle Columbia Mission STS-5 in November
1982.

179
 REFERENCES

1. Butow, Brig. Gen. Steven J., et al. State of the Space Industrial Base 2020,

United States Space Force, Jul. 2020.

2. Mattis, Jim. "Summary of the National Defense Strategy of the United States of

America." U.S. Department of Defense, Jan. 2018.

3. Harrison, Todd, et al. Space Threat Assessment 2019, The Center for Strategic

and International Studies, Apr. 2019.

4. Garamone, Jim. "Esper: Air Force, Space Force Leading Charge to New

Technologies." DOD News, U.S. Department of Defense, 16 Sept. 2020,

https://www.defense.gov/Explore/News/Article/Article/2349408/esper-air-force-

space-force-leading-charge-to-new-technologies/

5. Erwin, Sandra. "Raymond calls out Russia for 'Threatening Behavior' in Outer

Space." SpaceNews, SpaceNews, 12 Feb. 2020,

https://spacenews.com/raymond-calls-out-russia-for-threatening-behavior-in-

outer-space/

6. National Academies of Sciences, Engineering, and Medicine. 2016. Achieving

Science with CubeSats: Thinking Inside the Box. Washington, DC: The National

Academies Press. doi:10.17226/23503.

7. California Polytechnic SLO, CubeSat Design Specification. San Louis Obispo,

CA: California Polytechnic

180
8. NASA CubeSat Launch Initiative, CubeSat 101: Basic Concepts and Principles

for First-Time CubeSat Developers. NASA, Oct. 2017.

9. Reed, Helen L., Final Report: NNJ06HG15G AggieSat2: Autonomous

Rendezvous and Docking, Texas A&M Engineering Experiment Station, Dec.

2016.

10. Fury, Kris. "Launching Traffic Cameras Into Space." Lawrence Livermore

National Laboratory Science & Technology Review, Apr. 2012, pp. 4-10.

11. Wertz, James R, David F. Everett, and Jeffery J. Puschell. Space Mission

Engineering: The New SMAD. Hawthorne, CA: Microcosm Press, 2011. Print.

12. “STARE (Space-Based Telescopes for the Actionable Refinement of

Ephemeris).” STARE - EoPortal Directory - Satellite Missions,

https://directory.eoportal.org/web/eoportal/satellite-missions/s/stare.

13. Vincent Riot, Willem de Vries, Lance Simms, Brian Bauman, Darrell Carter,

Don Phillion, Scot Olivier, “The Space-based Telescopes for Actionable

Refinement of Ephemeris (STARE) mission,” Proceedings of the

27th AIAA/USU Conference, Small Satellite Constellations, Logan, Utah, USA,

Aug. 10-15, 2013, paper: SSC13-XI-11.

14. NASA, NASA-ESPC-2903-B JEM Payload Accommodation Handbook, NASA,

Mar. 2014.

15. NASA, JSC 27301F Materials and Processes Selection, Control, and

Implementation Plan for JSC Flight Hardware, NASA, Aug. 2009.

181
16. NASA, SSP 50835D ISS Pressurized Volume Hardware Common Interface

Requirements Document, NASA, Apr. 2013.

17. Destain, Adelin. AGS4-05-05-004-R0 Structural Analysis Report, AggieSat

Lab, 15 Jul. 2015.

18. NASA, "Cargo_Tool_revK-Dragon Environment.xlsx", JSC Engineering

Technology & Science Contract, March, 2015.

19. NASA, NASA-STD 8739.3 Soldered Electrical Connections, NASA, 6 Jun. 2008.

20. Cooper, Jake. AGS4-05-05-006-R0 Thermal Analysis Report, AggieSat Lab, 16

Jun. 2015.

21. Graves, John Thomas, et al. "AggieSat2 Student Satellite Mission." 50th AIAA

Aerospace Sciences Meeting including the New Horizons Forum and Aerospace

Exposition, AIAA, 9-12 Jan. 2012, AIAA 2012-0434.

22. Technologic Systems, TS-7800, Technologic Systems Inc., 5 Feb. 2021,

https://docs.embeddedarm.com/TS-7800

23. Kim, Bong Tae, Rechargable Lithium Ion Battery Model: ICR18650 C2

2800mAh, LG Chem, 10 Apr. 2014.

24. Spectrolab, Ultra Triple Junction, Spectrolab, 16 Jul. 2014.

25. Graves, John Thomas. Small Satellite Applications of Commercial Off the Shelf

Radio Frequency Integrated Circuits, Texas A&M University, Dec. 2011.

26. Kenwood Corporation, TH-D72A Manual, Kenwood Corporation, 15 May 2011.

27. Texas Instruments, CC1101 Low Power Sub-1 GHz RF Transceiver, Texas

Instruments, Nov. 2011.

182
28. Stealth Microwave, SM04093-36HS, Stealth Microwave Inc. Jul. 2014.

29. Digi International, 9Xtend -PKG-R RS232/422/485 RF Modem, Digi

International, Mar. 2007.

30. NewSpace Systems, Sun Sensor Datasheet, NewSpace Systems, 2020.

31. VectorNav, VN-100 User Manual, VectorNav, 2015.

32. Sinclair Interplanetary, Microsatellite Reaction Wheels (RW3-0.06), Sinclair

Interplanetary, 2019.

33. Davis, Jason. LightSail update: Of booms and Pretty Pictures, The Planetary

Society, 1 Jul. 2014.

34. Perez, Joe. AGSL-3108-200-001Quality Management System, AggieSat Lab, 27

Mar. 2006.

35. NASA, NASA-STD-6001B Outgassing Standards, NASA, 26 Aug. 2011.

36. Tucker, Andrew L. AGSL-3105-200-001 Staking and Coating Procedure Rev A,

AggieSat Lab, 30 Jun. 2014.

37. NASA, NASA-STD-8739.1A Workmanship Standard for Polymeric Application

on Electric Assemblies, NASA, 4 Mar. 2008.

38. Tucker, Andrew L. AGSL-3105-200-002 Thread Locking Compound Procedure,

AggieSat Lab, 18 Aug. 2014.

39. NASA, NSTS 08307 Revision A Criteria for Preloaded Bolts, NASA, 6 Jul.

1998.

40. Davidoff, Martin. The Radio Amateur’s Satellite Handbook, 4th ed., The

American Radio Relay League, 2003.

183
41. U.S Department of Commerce, Man-Made Noise Power Measurements at VHF

and UHF Frequencies, Department of Commerce NTIA Report 02-390, 2001.

184
 APPENDIX A

LINK BUDGET COMPONENTS AND EQUATIONS

Units

This link budget is based in the decibel milliwatt (dBm). Transmit power in

Watts is converted into the logarithmic dBm, and all gains and losses of the system are

added. The final result, the signal-to-noise ratio (Eb/No), tells how much signal strength

is left once all processing has taken place, and indicates how well the transmissions will

go through. Power is converted into dBm through the equation:

( 1)

EIRP

The equivalent isotropic radiated power (EIRP) is how much power the signal

has after the initial gains and losses from all components (in dBm) between the transmit

radio and antenna have been added together. It is calculated through the following

equation:

( 2)

This number is the starting power when the signal leaves the antenna. System

loss was estimated [11], while pointing losses were calculated as explained in the

following section. Gain values are determined by the selected hardware.

185
Pointing Loss

Pointing loss occurs when the main lobe of the antenna radiation pattern does not

point at the receiving antenna. There is an angle called the half power beamwidth

(HPB), at which the loss is 3 dB. For monopoles, this is roughly 45o, 21o for Yagi's, and

80o for patch antennas [11]. The equation to calculate the pointing loss is [11] :

( 3)

The HPB is the “cone” inside which the signal is still above half strength. The

actual pointing offset at which the signal drops to half power is actually 0.5*HPB.

Pointing error for the ground station antenna rotor was measured to be approximately 4o,

which yields a ground station pointing loss of 0.44 dB.

The current mission plan is to use the ADCS torque coils to prevent the

spacecraft from wildly tumbling through its orbit, and keep the antennas at an orientation

suitable for ground passes. While not as efficient as actively pointing with the reaction

wheels, it is believed from experience with magnetic control on AGS2 that adequate

spacecraft control can be maintained for ground communication, though some pointing

loss will be incurred. The 20o used for spacecraft pointing loss in the budgets below

comes from an initial estimate of pointing ability from the ADCS team.

The pointing loss from the LDR and beaconing will likely be significantly higher

when not performing an HDR pass. For monopoles, depending on the pointing error,

pointing losses can range from 0-48 dB, with 48 dB occurring at a pointing error angle

186
of 90o. With the current setup, the UHF downlink can tolerate a pointing error of up to

47.6o before the link margin drops to 0 dB. Similarly, the VHF uplink system link

margin drops to 0 dB when the pointing error is 66.5o. These values are calculated using

the signal margin at 10o elevation, and as the spacecraft range decreases through the

pass, the signal strength increases and more pointing error is tolerable.

Free Space Loss

The propagation loss is the combined loss from free space losses (the natural loss

in signal strength over distance) and atmospheric and precipitation losses. Free space

loss is calculated through the following equation [11]:

( 4)

where R is the link range in meters, and λ is the signal wavelength in meters. In the

above budgets, the link range used the distance from the ground station to a point 10

degrees above the horizon at the mean altitude of the ISS (400 km). This is an estimate

for the furthest feasible distance for communication with the satellite due to line of sight

limitations. The effective received power (ERP) is the value of the signal strength after

propagation loss.

Atmospheric and Precipitation Loss

Atmospheric losses were calculated with the help of [11] through the equation:

( 5)

where β is the elevation angle and 0.3 is the loss in dB at 90 degrees elevation. Loss due

to precipitation, or moisture in the atmosphere, is based on numbers from [11].

187
Polarization Loss and Antennas

Total polarization loss occurs when a linearly polarized transmitter and receiver

are oriented at 90o with respect to each other. When they are at this orientation, the loss

is practically infinite because no part of the sinusoidal shaped signal can enter the

receiver. When their orientation is at 0o, there is no polarization loss, and the loss varies

with their orientation angle. A circular to linear scheme, and vice versa, avoids these

problems, since the signal always has an appropriate path to follow on the receiving

antenna. To avoid the possibility of this infinite loss, and to remove the need for attitude

control in one axis of stabilization, circularly polarized antennas were selected for the

ground station. Circular polarization allows the signal to travel essentially in a

corkscrew pattern, which means that some portion of the signal will always enter the

receiver no matter the orientation angle. Because of this, the maximum polarization loss

is 3 dB, and is reflected in the link budgets [40].

To get this circular polarization pattern and a high gain, a Yagi antenna with

elements positioned at 90o to each other will be used for all systems. M2 Antenna makes

such products, and their model 436CP42UG offers 18.9 dB gain for UHF frequencies.

This antenna was used for the AGS2 mission, was already installed at the Riverside

ground station, and will be used for the HDR downlink because of its high gain. The M2

2MCP14 was selected for the LDR VHF antenna because of its short length (to avoid

guy wire interference) that delivers 12.3 dB gain. The M2 436CP30 was selected for the

LDR UHF system antenna since it is smaller in form factor than the HDR antenna, but

still provides adequate gain, 15.5 dB. A reason that all of these antennas were selected is

188
because they are circularly polarized. Three separate antennas were selected to avoid

switching problems and signal loss.

Receive Amplification

In calculating the gains and losses for each system, it became apparent that the

radio systems would need amplification to have good margin. Low noise amplifiers

(LNA) condition the signal such that the noise downstream of the amplifier is quieted so

the received signal is not dwarfed by noise losses. The Amplitech 00250050-0810-D4

offers 25 dB gain to the amateur UHF frequencies, and has been used by AggieSat Lab

in previous tests. This LNA will be used for the HDR and LDR UHF systems.

Noise Temperature

Noise temperature values detailed in this section feed into the noise power

calculations in the following section.

The first step in calculating total noise temperature is to calculate the receiver

noise figure (NF). The formula to calculate the NF is as follows [25]:

( 6)

where PBER is the decibel margin required for transmissions with 10-5 bit error rate

(BER), and the noise floor is calculated through the equation [25]:

( 7)

Noise bandwidth and its requirements are discussed in the following section.

The value -173.8 dBm is the noise temperature value of a 50 Ohm resistor at 290 K,

189
which approximates a standard radio receiver at room temperature, and the noise

bandwidth is a function of the data rate (as discussed in section 3.9 [25]).

The two equations above are solved to find the Rx NF in dB. With the Rx NF

known in dB, it is converted out of logarithmic scale by

( 8)

The non-logarithmic NF is then converted to noise temperature through the noise

figure equation [11]:

( 9)

where TR is the receiver noise temperature, and To is the reference temperature of 290 K.

The noise temperature for the low noise amplifier (if used) is calculated in the same

manner (with the manufacturers NF). The system noise temperature for a single

amplifier is the given by the Friis equation stated below [11]:

( 10)

This temperature is then converted back into the system NF through the noise

figure equation. Note: This value will be in non-logarithmic form.

The ambient NF value is converted to non-logarithmic form, and then is added to

the system NF. This total NF can then be converted to temperature through the noise

figure equation. This total noise temperature is the effective temperature, and is used in

calculating the total noise power (section 3.9 [25]).

190
 In the calculation of link margin, the ambient NF used for the ground station is

that of a residential area (3 dB) [41]. Due to its location away at the Riverside Campus,

it is believed that the NF will actually be closer to that of a rural setting (-3 dB) [41].

Crosslink assumes the average noise temperature of the Earth, 290 K (3 dB NF) [11], is

being seen by the crosslink antenna.

Noise Power

The noise power is the effect of losses due to system components and how they

operate. The noise power (No) is calculated through the equation:

( 11)

where k is the Boltzmann constant (1.38*10^-23 J/K), B is the signal bandwidth (Hz), T

is the system noise temperature in Kelvin. The Boltzmann constant is multiplied by

1000 to convert into dBm, which allows the units to match up with those used in the

transmit portion of the link budget. Because the VHF and UHF systems operate on

AFSK and FSK respectively, their noise bandwidth is twice the data rate [11].

Signal to Noise Ratio Calculations

First, the transmit power is converted to decibel-milliwatts, and used to calculate

the EIRP. Then the propagation loss is calculated, and subtracted from the EIRP. The

receive system gains and losses are added and subtracted to obtain the effective received

power (ERP). The signal-to-noise ratio (Eb/No) is defined as:

( 12)

191
 The Eb/No’s required to ensure a bit error rate of 10-5 for each system were

obtained from [11] based on the modulation scheme, and are listed in the budgets below.

The system margin is how much signal strength is left over once the required Eb/No is

subtracted from the received Eb/No. The higher the margin, the better the signal

strength will be, and the likelihood of transmission errors is lower.

Noise Temperature Calculations

Table 32: LDR VHF Noise Temperature

LDR VHF

Receiver Noise Calculation

Sensitivity = P(BER) + Rx NF + Noise

Floor

Sensitivity -121.00 dBm

Noise Floor -140.00 dBm
Req. P(BER) 13.30 dB

Rx Noise Figure 5.70 dB

Rx NF
(dB) Rx NF Tr (K) Gain

Receiver 5.70 3.713546 786.9282

System Noise Temperature 786.93 K

System Noise Figure 3.71

Residential Noise Figure 3.00 dB

2.00

Total Noise Figure 4.71

Total Noise Temperature 1075.55 K
192
 Table 33: LDR UHF Noise Temperature
LDR UHF

Receiver Noise Calculation

Sensitivity = P(BER) + Rx NF + Noise

Floor

Sensitivity -121.00 dBm

Noise Floor -130.97 dBm
Req. P(BER) 13.30 dB

Rx Noise Figure -3.33 dB

Rx NF Gain
(dB) Rx NF Tr (K) (dB)

LNA 0.80 1.202264 58.65669 25

Receiver -3.33 0.464193 -155.384

System Noise Temperature 52.44 K

System Noise Figure 1.18

Residential Noise Figure 3.00 dB

2.00

Total Noise Figure 2.18

Total Noise Temperature 341.07 K

193
 Table 34: HDR Noise Temperature
HDR

Receiver Noise Calculation

Sensitivity = P(BER) + Rx NF + Noise

Floor

Sensitivity -95 dBm

Noise Floor -118.93 dBm
Req. P(BER) 13.3 dB

Rx Noise Figure 10.63 dB

Rx NF Gain
(dB) Rx NF Tr (K) (dB)

LNA 0.8 1.202264 58.65669 25

Receiver 10.63 11.54992 3059.475

System Noise Temperature 181.04 K

System Noise Figure 1.62

Residential Noise Figure 3.00 dB

2.00

Total Noise Figure 2.62

Total Noise Temperature 469.66 K

194
 Table 35: Crosslink Noise Temperature
Crosslink

Receiver Noise Calculation

Sensitivity = P(BER) + Rx NF +
Noise Floor

Sensitivity -110 dBm

Noise Floor -130.97 dBm
Req. P(BER) 13.30 dB

Rx Noise Figure 7.67 dB

Rx NF
(dB) Rx NF Tr (K) Gain

Receiver 7.67 5.843846 1404.715

System Noise Temperature 1404.72 K

System Noise Figure 5.84

Earth Background Noise Figure 3.00 dB

2.00

Total Noise Figure 6.84

Total Noise Temperature 1693.34 K

195
 APPENDIX B

CUSTOM OMNETICS WIRE HARNESS EXAMPLE

196
 APPENDIX C

FASTENER TORQUE CALCULATION

This appendix details how threaded fastener torque levels were calculated for

AGS4, and is based on material from the NASA standard NSTS-08307 Criteria for

Preloaded Bolts [39]. This section provides an example for calculating the torque

necessary to hold the battery boxes to the AGS4 bus.

Variables

Table 36: Fastener Properties

Variables Name Example Note
TPI Threads Per #4-40 has 40 threads per
Inch inch
D Major Diameter 0.112” for #4-40
TP Thread Pitch 1/TPI
Kb Bolt Stiffness 11000 ksi for Stainless Fastener material property
(Shear Strength) Steel
Eb Young’s 2.92*107 psi for A-286 Fastener material property
Modulus Steel
αb Coefficient of 9.11*10-6 /°F for A-286 Fastener material property
Thermal Steel
Expansion
UTS Ultimate 87000 psi for A-286 Fastener material property
Tensile Strength Steel
At Tensile Area 0.00616in2 for #4-40 π/4*(D-0.938194*TP)2
PAt Max. Bolt Load UTS * At

Table 37: Heli-Coil Properties

Variables Name Example Note
Lhc Heli-Coil 0.224” for an MS21209- Found in Heli-Coil
Length C0420 manual
Dpmin_ext Pitch Diameter 0.1283” for an MS21209- Found in Heli-Coil
C0420 manual
TP Locking 48oz-in for #4-40 Heli- Found in Heli-Coil
Torque Coil manual
197
 Table 38: Joint Properties
Variables Name Example Note
Fsu Ultimate Shear Aluminum 6061-T6 Joint material property
Strength 30000psi
nL1 Force Application specific 1/2 of each pinched
Application material
Point Distance
L1 Total Joint Application specific Pinched material + Thread
Thickness engagement length
Kj Joint Stiffness 3770 ksi for 6061-T6 Joint material property
(Shear
Strength)
Ej Young’s 1.00*107 psi for 6061-T6 Joint material property
Modulus
αj Coefficient of 1.31*10-5/°F for 6061-T6 Joint material property
Thermal
Expansion

Calculating External Loads

The external loads that any fastener on an assembly could experience should be

determined from the root mean square acceleration load the mass being fastened will

experience, in this case that of the launch vehicle. This resultant force is then divided

over the minimum number of fasteners that secure the joint.

Load Calculation Formula/Method

The external load P(lbf) is calculated through the following equation:

( 13)

m - Assembly mass (kg)

g - Earth gravitation acceleration (9.8 m/s2)

Grms - Root mean square acceleration of launch vehicle, (3.2 g’s)

198
 Ns – Number of screws (dimensionless)

The "3" in the above equation represents the 3σ distribution from the SpaceX

PSD curve, in order to account for the statistical highest loads that could be seen (~ 9

g's).

Thermal Loads

Thermal loads depend on the temperature extremes the spacecraft will

experience, as well as the temperature at which the bolts were fastened into the

assembly. Temperature change will cause the fastener and joint to expand or contract,

increasing or decreasing the load on the bolt. In order to determine the temperature

extremes experienced on orbit, a basic thermal analysis must be performed. These

temperature changes are used to determine how much the bolt load changes.

Thermal Analysis

The thermal analysis performed for AGS4 was modified to calculate the effects

of a worst case scenario, with AGS4 directly between the Earth and Sun, with the Sun

shining directly on the -Z panel where the batteries are mounted, and the reflections from

Earth striking the +Z panel. Panels were allowed to radiate to the Earth and free space

as described in the thermal analysis, and the internal components and panels were

allowed to radiate to each other. To model the worst case scenario where thermal

stresses are highest, this configuration was allowed to reach steady state conditions.

SolidWorks allows the user to probe locations in the model to find the calculated

temperature at that point.

199
 Figure 112: Battery Box Thermal Analysis Probing

Data of immediate usefulness to bolt load calculations is the total temperature the

fasteners in the assembly will reach. Probing the results determined that the maximum

temperature reached by the boxes and fasteners was 57 °C (134.6 °F). This value is

within the survivable range for the batteries and operation. The worst cold case from the

dynamic thermal analysis calculated that the lowest temperature these boxes should

experience is -9.7 °C (14.5 °F).

Thermal Preload Calculation

An important factor in determining thermal bolt loads is the temperature at which

the fastener was installed. They should be at the same temperature of AGSL at

200
installation, which was roughly 70°F (21°C), approximately room temperature. Since

NSTS 08307 uses U.S. customary units, these calculations will be in that system as well.

( 14)
( 15)

These temperature differences are used in the following equations to find the

positive and negative thermal loads respectively.

( 16)
( 17)

L1 and L2 represent the thickness of materials being clamped, and are depicted in

the figure below. Both have a value of ¼” for the battery boxes.

Figure 113: Fastener Dimension Diagram

Using the above equations and the material properties listed in Table 36, Table

37, and Table 38, the resultant thermal loads are as follows:

201
 Table 39: Battery Box Thermal Preload
Thermal Preload
L1 0.25 in.
L2 0.25 in.
pos
Pthr 46.35 lbf
neg
Pthr -39.82 lbf

Calculating Preload

Based on the thermal loads, major diameter of the bolt, uncertainty coefficients,

and arbitrary torque values, the preload minimum and maximum values can be

determined. A spreadsheet was used to calculate these values over a wide range of

torque values according to the following equations, as preload depends on the applied

torque. This method will prove useful for calculating other parameters detailed in

following sections.

( 18)

( 19)

( 20)

where:

PLDmax - Maximum preload (lbf)

PLDmin - Minimum preload (lbf)

Ploss - Preload loss

Uncertainty factor (0.35 used for unlubricated bolts)

202
 Ktyp - Nut factor – generally 0.2 for most applications

T - Applied torque value

Table 40: Uncertainty Factor Γ

Torque Measurement of Unlubricated Bolts ±35%
Torque Measurement of Cadmium Plated ±30%
Bolts
Torque Measurement of Lubricated Bolts ±25%
Hydraulic Tensioner Installation ±15%
Preload Indicating Washers ±10%
Ultrasonic Measurement Devices ±10%
Bolt Elongation Measurement ±5%
Instrumented Bolts ±5%

Spreadsheet Calculation

A spreadsheet was created that listed torque values incremented in tenths of an

in-lb ranging from 1.0 to 10.0. Other values should be attempted depending on the

application. The formulas for maximum, loss, and preload were used with each torque

value to calculate a maximum and minimum preload at that torque. The margin of safety

(see subsequent section) was calculated for each of these preloads to determine the

acceptable range of torque values for installation of the bolts.

Hole Shear Out Force Calculations

Knowing that the ultimate shear strength of aluminum is 30000 psi, a rough but

conservative estimate for the maximum hole shear out strength (PAs) is as follows:

( 21)
where Le is the length of thread engagement, which is determined by:

( 22)

203
 The 1.5 factor in the above equation is due to the length of the locking feature of

the Heli-Coil (1 thread width), and the distance from the top of the tapped hole to the

Heli-Coil threads (0.5 thread width). Adding the values from both of these factors

yields the resultant 1.5 threads.

Using the above equations for a #4-40 screw and Heli-Coil, and the material

properties listed in Table 36, Table 37, and Table 38, the hole shear out strength is

calculated.

Margin of Safety

Two tensile margin of safety (MS) criteria must be met for each bolt in the

assembly:

( 23)
( 24)

where SF is the desired factor of safety, P is the external load, and PAt is the maximum

allowable tensile strength of the bolt. A SF = 3 was chosen to ensure that the load could

be handled. Pb, the axial bolt load is determined by:

( 25)

( 26)

( 27)

204
 In the spreadsheet described above, additional columns were added for both

margin of safety criteria and the value of Pb. These also change with the incrementing

torque values, with the exception of the MST1, which is constant with torque.

In addition to the tensile margin of safety, two hole shear out margin of safety

criteria must be met for each joint in the assembly:

( 28)
( 29)

Calculation and Results

The above equations are used to calculate the following results for the fastener

joint parameters.

Table 41: Joint Calculation Results

Battery Box Mass M 2.9 (1.3) lbf (kg)
Number of Fasteners NS 14
External Load (per fastener) P 1.96 lbf
pos
Positive Thermal Load Pthr 46.35 lbf

Negative Thermal Load Pthr neg -39.82 lbf

Hole Shear out strength PAs 761 lbf

Tensile Margin of Safety-1 MST1 89.9 lbf
Shear Margin of Safety-1 MSS1 128.2 lbf

With the above quantities calculated, the preload, axial bolt force, and two

remaining margins of safety for each fastener torque value can be calculated. These

quantities are investigated over a range of torque values to find a value where both

remaining margin of safety values are greater than zero, and both PLDmax and PLDmin

205
are positive. The range of acceptable values often encompasses several in-lb. From

personal experience in automotive applications, a value towards the upper third of

acceptable values is usually chosen as it helps to make sure the fastener stays tight, while

leaving margin. The table below shows some of the torque values investigated and the

associated margins. A value of 7 in-lb was selected as it meets the criteria mentioned

above.

Table 42: Preload and Margin of Safety for Fastener Torque

Torque
Value
(in*lbs) PLDmax Ploss PLDmin Pb MST2 MSS2
4 287.4169 14.37085 -10.661 288.267781 0.858414498 1.640229959
5 347.6848 17.38424 15.34344 348.535638 0.537062397 1.183688405
5.5 377.8187 18.89094 28.34567 378.669567 0.414745389 1.009913916
6 407.9527 20.39763 41.3479 408.803495 0.310460967 0.861758132
6.5 438.0866 21.90433 54.35013 438.937424 0.220495211 0.733944727
7 468.2205 23.41103 67.35237 469.071352 0.142088556 0.622553217
7.5 498.3544 24.91772 80.3546 499.205281 0.073147749 0.524609736
8 528.4884 26.42442 93.35683 529.33921 0.01205619 0.4378176
8.5 558.6223 27.93111 106.3591 559.473138 -0.042454432 0.360374931
9 588.7562 29.43781 119.3613 589.607067 -0.091393143 0.290848219

206