Thermal Vacuum Chamber Modification Testing and Analysis

THERMAL VACUUM CHAMBER MODIFICATION, TESTING, AND
ANALYSIS
A Thesis
presented to
the Faculty of California Polytechnic State University,
San Luis Obispo
In Partial Fulfillment
of the Requirements for the Degree
Master of Science in Aerospace Engineering
by
Jared Lehmann
July 2021
© 2021
Jared Lehmann
ALL RIGHTS RESERVED
ii
COMMITTEE MEMBERSHIP
TITLE: Thermal Vacuum Chamber Modification,
Testing, and Analysis
AUTHOR: Jared Lehmann
DATE SUBMITTED: July 2021
COMMITTEE CHAIR: Kira Abercromby, Ph.D.
Professor of Aerospace Engineering
COMMITTEE MEMBER: Pauline Faure, Ph.D.
Assistant Professor of Aerospace Engineering
COMMITTEE MEMBER: Jennifer Mott, Ph.D.
Assistant Professor of Mechanical Engineering
COMMITTEE MEMBER: Kim Shollenberger, Ph.D.
Professor of Mechanical Engineering
iii
ABSTRACT
Thermal Vacuum Chamber Modification, Testing, and Analysis
Jared Lehmann
This work dissusses the modification and analysis of the Blue Thermal Vacuum Cham-
ber (TVAC) located at the Space Environments Lab at California Polytechnic State
University, San Luis Obispo. The modified design has a cylindrical test section and
can accommodate 6U Cubesats or larger for educational or research purposes.
The sizing process for the modified shroud cooling system and modular heating
plates is discussed. The modified cooling system uses existing nitrogen plumbing into
the chamber and control systems with a new copper shroud. The modified heating
system uses modular heater plates, which utilize the existing three heater strips.
The modified system includes high emissivity coatings for improved heat transfer
performance, lower thermal mass materials to minimize thermal mass and liquid
nitrogen consumption, and modular components for flexibility in operation.
Analysis presented shows correlation between experimental results and a steady
state thermal model using SolidWorks and SolidWorks Flow Simulation. The results
demonstrate a maximum absolute difference in modeled vs experimental temperatures
at measured locations of 11◦ C in all cases, and 3◦ C for test article temperatures only.
Chamber performance is compared and characterized through a series of thermal
vacuum tests and demonstrates capability exceeding ISO 19683 requirements for all
thermal vacuum chamber testing categories except tolerance, with a tested tempera-
ture range of -145◦ C at the shroud to 95◦ C at the heater plates, >10 cycles between
-15◦ C and 55◦ C, dwells in excess of 3 hours, ramp rates of 1-2◦ C/min, and chamber
pressures under <7.5x10−6 Torr after bakeout procedures.
iv
ACKNOWLEDGMENTS
I would like to thank my parents for their support throughout my education and
my life and my girlfriend, Brooke, who supported me through my thesis.
To Dr. Abercromby, you have taught me so much and have been such a great
part of my college education, having taken at least 5 courses from you and having
you be my thesis advisor. You always have great advice and I have learned so much
from you.
To Dr. Faure, thank you for your guidance and support throughout classes, senior
design, and my thesis.
To Cody Thompson, I cannot thank you enough for your assistance, advice, and
knowledge about manufacturing. Without you this project would not have been
possible.
Thank you to Dr. Shollenberger and Dr. Mott, for being apart of my thesis
committee and for your support and feedback along the way.
v
TABLE OF CONTENTS
Page
LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 The Thermal Space Environment . . . . . . . . . . . . . . . . . . . . 2
1.3 Thermal Testing and Standards . . . . . . . . . . . . . . . . . . . . . 4
1.4 Thermal Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4.1 Fundamental Equations . . . . . . . . . . . . . . . . . . . . . 6
1.4.2 SolidWorks vs Thermal Desktop . . . . . . . . . . . . . . . . . 7
1.5 Experimental Apparatus Background and Previous Capabilities . . . 10
1.6 Limitations and Modifications . . . . . . . . . . . . . . . . . . . . . . 13
1.7 Thesis Motivation and Work . . . . . . . . . . . . . . . . . . . . . . . 15
2 Modification Design Development . . . . . . . . . . . . . . . . . . . . . . . 18
2.1 Goals and Design Criteria . . . . . . . . . . . . . . . . . . . . . . . . 18
2.2 Cooling Shroud Sizing . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2.1 Initial Configuration Comparisons . . . . . . . . . . . . . . . . 19
2.2.2 Detailed Configuration Comparison . . . . . . . . . . . . . . . 22
2.2.2.1 Numerical Model Setup . . . . . . . . . . . . . . . . 22
2.2.2.2 Single and Double Spiral Comparison . . . . . . . . . 23
2.2.3 Single Spiral Optimization . . . . . . . . . . . . . . . . . . . . 27
2.2.4 Configuration Selection . . . . . . . . . . . . . . . . . . . . . . 31
2.3 Heater Plate Sizing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.3.1 Heater Plate Thermal Model . . . . . . . . . . . . . . . . . . . 33
2.4 High Emissivity Coatings . . . . . . . . . . . . . . . . . . . . . . . . . 34
2.5 Platen, Modular Mounting System . . . . . . . . . . . . . . . . . . . 35
2.6 Aluminum Foil Reflectors . . . . . . . . . . . . . . . . . . . . . . . . 36
2.7 Final Modification Design . . . . . . . . . . . . . . . . . . . . . . . . 36
2.8 Liquid Nitrogen Tank Upgrade . . . . . . . . . . . . . . . . . . . . . . 37
vi
3 Manufacturing Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.1 Shroud Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.1.1 Copper Tubing and Sheet Properties . . . . . . . . . . . . . . 38
3.1.2 Tubing Forming . . . . . . . . . . . . . . . . . . . . . . . . . . 39
3.1.3 Shroud Assembly Brazing . . . . . . . . . . . . . . . . . . . . 40
3.1.4 Shroud Assembly Painting Coupon Testing . . . . . . . . . . . 46
3.1.5 Shroud Assembly Painting . . . . . . . . . . . . . . . . . . . . 50
3.2 Heater Plate Machining . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.3 Modular Platen Assembly . . . . . . . . . . . . . . . . . . . . . . . . 57
3.4 Outgassing Considerations . . . . . . . . . . . . . . . . . . . . . . . . 59
3.5 Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.6 Cost and Schedule . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4 Testing and Analysis Methodology . . . . . . . . . . . . . . . . . . . . . . 67
4.1 Thermal Vacuum Chamber Testing Configuration . . . . . . . . . . . 68
4.1.1 Chamber Bakeout . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.1.2 Controller Configuration . . . . . . . . . . . . . . . . . . . . . 68
4.1.3 PID Tuning . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.1.4 Data Collection . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.1.5 Thermocouple Placement . . . . . . . . . . . . . . . . . . . . . 72
4.1.6 Initial Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
4.1.6.1 Control Thermocouple Placement . . . . . . . . . . . 72
4.1.6.2 Coolant Leak Repair . . . . . . . . . . . . . . . . . . 73
4.1.6.3 Door Shroud Operation . . . . . . . . . . . . . . . . 73
4.1.7 Empty Test 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
4.1.7.1 Thermocouple Placement . . . . . . . . . . . . . . . 74
4.1.7.2 Experimental Configuration . . . . . . . . . . . . . . 75
4.1.7.3 Thermal Vacuum Test Profile . . . . . . . . . . . . . 75
4.1.8 Empty Test 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.1.9 CubeSat Mass Model Test . . . . . . . . . . . . . . . . . . . . 77
vii
4.1.9.4 Aluminum Plates Test . . . . . . . . . . . . . . . . . 81
4.2 Minimum Temperature Testing . . . . . . . . . . . . . . . . . . . . . 84
4.3 Thermal Model Setup . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.3.1 Assumptions and Simplifications . . . . . . . . . . . . . . . . . 85
4.3.2 Material and Surface Properties . . . . . . . . . . . . . . . . . 88
4.3.3 Mesh & Mesh Refinement Study . . . . . . . . . . . . . . . . . 90
4.3.4 Thermal Model Validation . . . . . . . . . . . . . . . . . . . . 91
5 Results and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.1 Leak Testing and Repair . . . . . . . . . . . . . . . . . . . . . . . . . 94
5.2 Empty Test 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
5.2.1 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . 96
5.2.2 Numerical Model Results . . . . . . . . . . . . . . . . . . . . . 98
5.2.2.1 Simulating Shroud Coolant Flow . . . . . . . . . . . 101
5.2.3 Model vs. Experimental Results Comparison . . . . . . . . . . 102
5.3 Empty Test 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105
5.4 CubeSat Mass Model Test . . . . . . . . . . . . . . . . . . . . . . . . 108
5.5 Aluminum Plates Test . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5.6 Thermal Cycling Consistency . . . . . . . . . . . . . . . . . . . . . . 116
5.7 Thermal Model Correlation Summary . . . . . . . . . . . . . . . . . . 117
6 Conclusion and Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . 119
viii
6.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
6.1.1 Temperature Range . . . . . . . . . . . . . . . . . . . . . . . . 120
6.1.2 Chamber Pressure . . . . . . . . . . . . . . . . . . . . . . . . 120
6.1.3 Number of Cycles . . . . . . . . . . . . . . . . . . . . . . . . . 121
6.1.4 Tolerance Limits and Temperature Variation . . . . . . . . . . 121
6.1.5 Thermal Dwell . . . . . . . . . . . . . . . . . . . . . . . . . . 122
6.1.6 Temperature Ramp Rate . . . . . . . . . . . . . . . . . . . . . 122
6.1.7 Test Article Volume . . . . . . . . . . . . . . . . . . . . . . . 123
6.1.8 Thermal Model Correlation . . . . . . . . . . . . . . . . . . . 123
6.2 Lessons Learned . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
6.2.1 Modification Design . . . . . . . . . . . . . . . . . . . . . . . 124
6.2.2 Manufacturing Processes . . . . . . . . . . . . . . . . . . . . . 125
6.2.3 Thermal Vacuum Testing . . . . . . . . . . . . . . . . . . . . 125
6.3 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
6.3.1 Powered Test Article . . . . . . . . . . . . . . . . . . . . . . . 126
6.3.2 Independent Heater and Shroud Control Tests . . . . . . . . . 126
6.3.3 Leak Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
6.3.4 Heater Modifications . . . . . . . . . . . . . . . . . . . . . . . 127
6.3.5 Cooling System Improvements . . . . . . . . . . . . . . . . . . 128
6.3.6 Data Acquisition System . . . . . . . . . . . . . . . . . . . . . 129
6.3.7 Cooling System Insulation . . . . . . . . . . . . . . . . . . . . 130
6.3.8 Webcam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
APPENDICES
A Bakeout Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
B Updated Chamber Operating Procedure . . . . . . . . . . . . . . . . 138
B.1 Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
B.2 Chamber Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
B.3 Vacuum System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
B.4 Heating System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
B.5 Cryogenic Cooling System . . . . . . . . . . . . . . . . . . . . . . . . 139
B.6 Operating Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
ix
B.6.1 Starting The Cooling System . . . . . . . . . . . . . . . . . . 139
B.6.2 During Normal Operation . . . . . . . . . . . . . . . . . . . . 140
B.6.3 Shutdown . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
B.6.4 Opening the Dewar and Ordering LN2 . . . . . . . . . . . . . 140
B.7 Contingencies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
B.7.1 O2 Sensor Alarm (Evacuation Procedure) . . . . . . . . . . . . 140
B.7.2 Increased Chamber Pressure . . . . . . . . . . . . . . . . . . . 140
B.7.3 Low Dewar Pressure (<3 PSIG) . . . . . . . . . . . . . . . . . 141
C Solidworks Flow Simulation Thermal Model Guide . . . . . . . . . . 142
C.1 Adding a Test Article . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
C.2 Setting Up A Simulation . . . . . . . . . . . . . . . . . . . . . . . . . 143
C.3 Meshing, Solving, and Visualizing Results . . . . . . . . . . . . . . . 146
x
LIST OF TABLES
Table Page
1.1 Previous Capability vs ISO 19683 Requirements for TVAC Testing 12

1.2 Rules and standards for thermal vacuum cycling tests from Chisabas
et al. [4] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.1 Cooling sizing analysis summary table, best configurations high-

lighted in green . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.1 Empty test 1 thermal vacuum test profile . . . . . . . . . . . . . . . 77

4.2 Empty test 2 thermal vacuum test profile . . . . . . . . . . . . . . . 78
4.3 CubeSat mass model thermal vacuum test profile . . . . . . . . . . 81
4.4 Plates thermal vacuum test profile . . . . . . . . . . . . . . . . . . 84
4.5 Chamber material properties for thermal model . . . . . . . . . . . 88
4.6 Chamber surface properties for thermal model . . . . . . . . . . . . 89
4.7 Mesh refinement study results . . . . . . . . . . . . . . . . . . . . . 90
5.1 Comparison of numerical model and experimental results for empty

test 1, 80◦ C hot case . . . . . . . . . . . . . . . . . . . . . . . . . . 103
test 1, -20◦ C cold case without flow simulation . . . . . . . . . . . . 104
test 1, -20◦ C cold case with flow simulation . . . . . . . . . . . . . 105
test 2, 55◦ C hot case . . . . . . . . . . . . . . . . . . . . . . . . . . 108
test 2, -20◦ C cold case without flow simulation . . . . . . . . . . . . 109
test 2, -20◦ C cold case with flow simulation . . . . . . . . . . . . . 110
5.7 Comparison of numerical model and experimental results for Cube-
Sat mass model, 80◦ C hot case . . . . . . . . . . . . . . . . . . . . 113
5.8 Comparison of numerical model and experimental results for the
plates test 90◦ C hot case . . . . . . . . . . . . . . . . . . . . . . . . 116
5.9 Comparison of maximum absolute difference between numerical model
and experimental results for all tests . . . . . . . . . . . . . . . . . 118
xi
LIST OF FIGURES
Figure Page
1.1 Spacecraft Energy Exchange Diagram [17] . . . . . . . . . . . . . . 3

1.2 Typical thermal vacuum test profile [7] . . . . . . . . . . . . . . . . 5
1.3 Thermal balance test profile [7] . . . . . . . . . . . . . . . . . . . . 6
1.4 Blue TVAC Chamber [11] . . . . . . . . . . . . . . . . . . . . . . . 10
1.5 Blue TVAC systems block diagram [11] . . . . . . . . . . . . . . . . 11
2.1 Concept shroud cooling tubing configurations . . . . . . . . . . . . 20

2.2 Double spiral temperature contour plot for the shroud with 2” spacing 25
2.3 Single spiral temperature contour plot for the shroud with 2” spacing 26
2.6 Single spiral temperature contour plot for the shroud with 3” spacing
and 3/8” OD tubing . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.7 Single spiral temperature contour plot for the shroud with 3” spacing
and 0.008” thick sheet . . . . . . . . . . . . . . . . . . . . . . . . . 31
2.8 Contour plot of heater plate front face temperature distribution . . 34
2.9 Final modification design section view with labeled components and
a CubeSat mass model test article shown . . . . . . . . . . . . . . . 36
3.1 Tubing cross section after rolling process . . . . . . . . . . . . . . . 39

3.2 1st brazing coupon with round tubing and thicker copper sheet . . 41
3.3 2nd brazing coupon with flattened tubing and 0.005” copper sheet . 42
3.4 Copper sheet during brazing process for cylinder forming . . . . . . 42
3.5 Copper coil roughly fit onto brazed sheet . . . . . . . . . . . . . . . 43
3.6 Shroud during brazing process . . . . . . . . . . . . . . . . . . . . . 44
3.7 Shroud in plastic tub during cleaning with dilute hot citric acid . . 45
3.8 Shroud after cleaning process with dilute hot citric acid . . . . . . . 46
3.9 One mounting rail test fit and bolted in place to determine bolt lengths 47
3.10 First primer coupon primer application with fisheyes and excessive
coating thickness . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
xii
3.11 Copper sheet coupon water break test before cleaning (upper) and
after cleaning (lower) . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.12 Second primer coupon with minimal fisheyes and even, thinner ap-
plication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.13 Paint test coupon after 2 coats of Aeroglaze Z306 on top of primer 51
3.14 Masked shroud after primer application . . . . . . . . . . . . . . . . 52
3.15 Shroud interior after Aeroglaze Z306 application . . . . . . . . . . . 53
3.16 Heater plate in the horizontal bandsaw to cut along the thickness . 54
3.17 Heater plate section before final machining . . . . . . . . . . . . . . 55
3.18 Heater plate back face after final machining . . . . . . . . . . . . . 56
3.19 Heater plate front face after final machining . . . . . . . . . . . . . 56
3.20 Heater plates after primer application with additional mounting holes
visible . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.21 Heater plates integrated into the shroud after painting . . . . . . . 58
3.22 Modular platen assembly installed on mounting rails inside shroud . 58
3.23 Shroud after mounting rails were installed . . . . . . . . . . . . . . 62
3.24 Trimmed shroud and chamber tubing after test fit . . . . . . . . . . 63
3.25 Shroud connector brazing setup with wet cooling towels . . . . . . . 64
3.26 Shroud coolant tubing connections after brazing and fastening . . . 64
3.27 Shroud and heater plates integrated into the chamber . . . . . . . . 65
4.1 Methodology overview flowchart . . . . . . . . . . . . . . . . . . . . 67

4.2 Auto-tuning process of the Watlow F4DH controller [21] . . . . . . 70
4.3 Labeled data collection experimental setup . . . . . . . . . . . . . . 71
4.4 Empty test 1 & 2 thermocouple placement diagram . . . . . . . . . 75
4.5 Empty test 1 & 2 configuration experimental setup . . . . . . . . . 76
4.6 Empty test 1 & 2 configuration experimental setup with front foil
reflector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.7 CubeSat mass model thermocouple placement diagram . . . . . . . 79
4.8 CubeSat mass model test article . . . . . . . . . . . . . . . . . . . . 79
4.9 CubeSat mass model test article setup in the chamber . . . . . . . 80
4.10 Plates used for testing, one bare aluminum, one painted black . . . 82
4.11 Plates thermocouple placement diagram . . . . . . . . . . . . . . . 83
xiii
4.12 Plate test articles setup in the chamber . . . . . . . . . . . . . . . . 83
4.13 Small mesh used for numerical simulations with conduction and ra-
diation only . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.14 Small mesh used for numerical simulations with conduction, convec-
tion (flow simulation), and radiation . . . . . . . . . . . . . . . . . 92
5.1 Vent valve leak source outlined in red . . . . . . . . . . . . . . . . . 95

5.2 Annotated chamber pressure during CubeSat mass model testing . 97
5.3 Logged temperature data during empty test 1 . . . . . . . . . . . . 98
5.4 Numerical model results for empty test 1, 80◦ C case . . . . . . . . . 99
5.5 Numerical model results for empty test 2, -20◦ C case without flow
simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
5.6 Numerical model results for empty test 2, -20◦ C case without flow
simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
5.7 Logged temperature data during empty test 2 . . . . . . . . . . . . 106
5.8 Numerical model results for empty test 2, 55◦ C case . . . . . . . . . 107
5.9 Logged temperature data during CubeSat mass model test . . . . . 111
5.10 Numerical model results for CubeSat mass model test, 80◦ C case . 112
5.11 Logged temperature data during aluminum plates test . . . . . . . 114
5.12 Logged temperature data during hot soak after aluminum plates test 114
5.13 Numerical model results for plates test, 90◦ C case . . . . . . . . . . 115
5.14 Annotated plot of absolute difference in CubeSat temperature re-
sponse for last 2 cycles of test . . . . . . . . . . . . . . . . . . . . . 117
6.1 Chamber modifications (right) shown with previous configuration

(left) [11] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
6.2 Insulation around the liquid nitrogen tank inlet and uninsulated ar-
eas with ice buildup . . . . . . . . . . . . . . . . . . . . . . . . . . 130
6.3 Insulation around chamber cryogenic valves and uninsulated areas
with ice buildup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
C.1 Adding an test article part or assembly to the thermal model . . . 142
C.2 Moving and mating the test article in the assembly . . . . . . . . . 143
C.3 Loading flow simulation and viewing cases and details . . . . . . . . 144
C.4 Specific parameter selection . . . . . . . . . . . . . . . . . . . . . . 145
C.5 Defining temperature boundary conditions for the hot case . . . . . 145
xiv
C.6 Defining temperature boundary conditions for the cold case, without
flow simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
C.7 Adding a radiative surface . . . . . . . . . . . . . . . . . . . . . . . 147
C.8 Adding a radiative surface in detail . . . . . . . . . . . . . . . . . . 147
C.9 Defining goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
C.10 Basic meshing using global and local meshes . . . . . . . . . . . . . 148
C.11 Running the mesher and solving the simulation . . . . . . . . . . . 149
C.12 Viewing the mesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
C.13 Viewing temperature surface plots and probing results . . . . . . . 150
xv
Chapter 1
INTRODUCTION
1.1 Background
Spacecraft are exposed to a hostile and varied environment, including constant and
cyclic bombardment of solar and electromagnetic radiation, neutral and charged par-
ticles, meteoroids, debris, and more [17]. Spacecraft must be designed to resist or
manage the effects of exposure to these environments. The thermal environment
plays a significant role in the design of spacecraft because the components and struc-
tures that make up the spacecraft have ranges of temperatures in which they can
survive and operate [7]. In order to maintain these temperatures in a vacuum en-
vironment, care must be taken in the design of thermal control systems to ensure
that these temperatures will be maintained throughout the mission. Mathematical
analysis is used to predict the performance of a spacecraft’s thermal control system
in a particular environment, but verification methods are needed to ensure that the
system meets design requirements.
A thermal vacuum (TVAC) chamber, used to conduct thermal vacuum testing,
can be utilized to perform verification to ensure the system meets design requirements.
To perform verification of the analytical models, the TVAC chamber must be able
to precisely control the thermal and vacuum environment as needed to meet the
requirements of the test and specification being followed. Typically, this includes
pumping the chamber down to a specified vacuum level, ramping the temperature
to a high temperature and holding for a period of time, then ramping down to a
cold case condition and holding for a period of time [4]. The hold period is called a
thermal soak or dwell. The intent of cycling between temperature extremes under high
1
vacuum is to verify the test article performs and meets requirements in a simulated
space environment [4]. The number of cycles and duration of soaks is specific to
requirements of each mission. However, the test levels are potentially limited by the
capabilities of the TVAC chamber used.
This thesis explores methods of improving the performance of the existing and
previously modified MDA US systems thermal vacuum chamber located in California
Polytechnic State University, San Luis Obispo’s space environments laboratory. Mod-
ifications will be made through cooling system and heater design improvements, the
addition of surface coatings, updated control schemes, and defined processes. After
modifications are made, the performance of the chamber will be analyzed through a
series of tests and compared to a software based numerical model using SolidWorks
and SolidWorks Flow Simulation.
1.2 The Thermal Space Environment
The thermal environment in space is dominated by radiative heat transfer because no
significant particles exist in the vacuum of space for effective convection or conduction
to the environment [17]. Any heat to or from the environment must be emitted or
absorbed and can include a variety of sources and sinks including direct solar radia-
tion, albedo, planetary radiation, and emitted radiation [17]. Direct solar radiation
is from the sun, and the heat flux for a blackbody is proportional to the spacecraft’s
distance from the sun [17]. Albedo is reflected radiation from the nearby celestial
body’s surface to the spacecraft [17]. Planetary radiation is the radiation emitted by
the celestial body due to the temperature of the body and is typically in the IR spec-
trum [17]. Emitted radiation from the spacecraft is the only method of transferring
heat to the environment [17]. These sources of radiation must be considered when
designing the thermal control systems for a spacecraft such that the incoming and
2
outgoing heat flux are generally balanced, depending on the specific requirements of
the mission. Figure 1.1 summarizes the sources and sinks between a spacecraft and
the environment.
Figure 1.1: Spacecraft Energy Exchange Diagram [17]
When designing a spacecraft’s thermal control systems, passive and active meth-
ods for thermal control can be used depending on the needs of the mission. Examples
of passive methods are coatings, which affect the emissivity and absorptivity of sur-
faces, can be applied as needed to dissipate or absorb energy [17]. MLI blankets
and thermal shields can be used to insulate interior faces from the external radiation
environment [17]. Thermal storage like phase change materials or cryogenics can be
used to regulate temperatures [17]. Heat pipes and radiators transfer and emit heat
passively [17]. Active methods of thermal control include resistance heaters, heat
pumps and pumped fluid loops to move heat around the spacecraft, thermoelectric
coolers to pump heat typically from small components, and active radiators to emit
heat selectively to the environment [17].
3
1.3 Thermal Testing and Standards
A thermal vacuum chamber can be used for a variety of test campaigns. Thermal
vacuum tests, thermal vacuum cycling, thermal balance testing, burn-in tests, and
thermal noise testing are among the types of tests that can be performed using a
thermal vacuum chamber [4, 6]. Each test has a specific purpose, but the overall
intent is to screen for environmental stress due to thermal expansion and cycling of
hot to cold temperatures, test the ability for the test article or its components to
turn on under specific thermal conditions, and to demonstrate that the test article
can survive under specific thermal conditions [4, 11, 6]. Any of these test objectives
can be evaluated or only some specific objectives can be evaluated, depending on
the requirements of the test campaign. The test article can be anything from an
individual component to a fully integrated satellite.
Thermal Vacuum Tests: This test verifies the performance and function of the
test article under vacuum conditions and a specified number of hot and cold cycles [7].
Figure 1.2 shows a typical thermal vacuum test profile. The number of cycles, soak
temperatures, margins, and ramp rates are all selected based on requirements, testing
capabilities, and relevant specifications. For example, NASA’s General Environmental
Verification Standard (GEVS) requires prototype qualification test temperatures to
be +10◦ C of the maximum expected flight temperature and -10◦ C from the minimum
expected flight temperature [1].
Thermal Balance Test: This test simulates flight conditions at the steady state
cold and hot conditions to verify thermal control systems and to correlate thermal
analytic models [7]. Figure 1.3 shows a simple thermal balance test profile. This test
is usually performed as part of the thermal vacuum testing for the system or satellite
[7].
4
Figure 1.2: Typical thermal vacuum test profile [7]
Thermal Cycle Test: This test verifies performance of the test article under rapid
cycling of hot and cold temperatures in an ambient or gaseous nitrogen environment,
which screens primarily for environmental stresses [7]. It should be noted that thermal
vacuum tests and thermal cycling tests typically use different test equipment due to
the different ambient environment and difference in heat transfer modes.
Thermal Burn-In Test: Verifies the test article is operational at a set temperature
or cycled for an added time as required. This test usually takes place as part of the
thermal cycling test [7]. Thermal noise can also be evaluated during a burn in test,
which is important for radiofrequency payloads [6].
1.4 Thermal Modeling
Thermal modeling is any mathematical tool used to predict temperatures, typically
on a system scale with different components and structures interacting with each
other and the environment [1]. This model is continuously developed and used to help
design a satellite thermal control system and is continuously updated and modified as
5
Figure 1.3: Thermal balance test profile [7]
a design progress. Eventually, the model can be used to evaluate the performance on
orbit and can be helpful to investigate anomalies or perform specialized operations or
maneuvers [12]. Software has made executing and analyzing thermal models simpler
and allows for a variety of analysis that hand calculations do not, including numerical
analysis. However, it is important to note that hand calculations are often helpful
before jumping into thermal modeling as a first pass analysis before more detailed
analysis can take place.
1.4.1 Fundamental Equations
Depending on the specific question, basic equations like those used for conduction,
convection, and radiation are below in equations 1.1,1.2, and 1.3 respectively. In these
equations, q̇ represents heat flux, typically measured in Watts per meter squared. In
equation 1.1, k refers to thermal conductivity and is typically measured in Watts per
meter-Kelvin, and ∇T refers to the temperature gradient in Kelvin per meter. This
6
equation can be used to model heat flowing through solids as well as between liquids
and gases via conduction.
q̇ = −k∇T [17] (1.1)
Equation 1.2 is similar to conduction except the coefficient h relates to the convec-
tion coefficient in Watts per meter squared-Kelvin, and this equation refers to heat
flowing due to fluid flow past a solid surface [17].
q̇ = −h∆T [17] (1.2)
Equation 1.3 represents radiative heat transfer, where eq is the emissivity, σ is
the Stefan-Boltzmann constant, and |Ti 4 − To 4 | is the absolute difference of the cube
of the inner surface temperature minus the cube of the inner surface temperature.
This equation is refers to radiative heat transfer between inner and outer surfaces,
and is directly proportional to emissivity.
q̇ = eq σ|Ti 4 − To 4 |[17] (1.3)
1.4.2 SolidWorks vs Thermal Desktop
For the analysis in this thesis, SolidWorks Simulation and SolidWorks Flow Simu-
lation was selected over continuing to use Thermal Desktop, used by Williams and
7
Jensma for their thermal modeling [11, 22]. The primary reason this change was
made was because of ease of use for Cal Poly students. As this chamber will be used
primarily by Cal Poly students, who learn computer aided design (CAD) in classes
using SolidWorks and have access to the software on their own computer or virtual/-
physical lab computers, using SolidWorks for the modeling and analysis reduces the
learning curve for users vs Thermal Desktop, which is generally unfamiliar to most
students.
The previous thermal models by Williams and Jensma used Thermal Desktop,
which uses SINDA/FLUENT [11, 22]. These software packages are generally more
flexible in that the user has more variety of models and solution methods to use, and
generally more options when it comes to setup and calculation. However, SolidWorks
as well as SolidWorks Flow Simulation has all of the same general capabilities and is
more than capable of analysis for this thesis, including flow simulation capabilities for
nitrogen through the coolant tubing as well as radiation ray tracing and conduction
finite element analysis. Thermal desktop is particularly useful in that it includes
specialized tools to determine thermal loads in specific orbits, where SolidWorks does
not. This is useful in some cases but for the purpose of thermal modeling of the
vacuum chamber, these thermal loads are not needed to be calculated and could
be manually imported into Solidworks based on a Thermal Desktop model or other
calculations to simulate specific heat loads. Additionally, in general SolidWorks is
more user friendly and more widely used, therefore documentation and online tutorials
are readily available which is not always the case for thermal desktop.
SolidWorks thermal and flow simulation uses a general-purpose thermal/fluid
solver which utilizes a finite volume method to solve partial differential equations
relating to fluid mechanics and heat transfer for convection and conduction [12, 20].
Starting from a solid model, the mesher prepares for analysis by breaking the as-
sembly into finite elements, like tetrahedrons. Governing thermodynamic equations
8
are used and integrated over each element. Each element has specific defined ther-
mal properties and initial conditions and contains multiple nodes, which are typically
located at the vertices of the finite element [12]. The equations are then solved nu-
merically for each element given the specific thermal properties and state of the node
and surrounding nodes [12]. For radiative heat transfer, SolidWorks utilizes a ray
tracing algorithm that creates paths from node surfaces to other node surfaces in or-
der to determine the view factors between all surfaces in the model [20]. This method
can be done forwards from source to sink or backwards from sink to source and is
called forward and backward ray tracing, respectively [20]. For steady state analysis,
each node is solved iteratively until the solution converges to a steady state within a
specified tolerance. In SolidWorks flow simulation, the results are able to be viewed
with the solid model and changed to allow for rapid iteration during the designing,
testing, and analysis process.
The intent of the thermal model is to predict how the chamber will perform so
that the response of the test article can be known or approximated before the test.
When the thermal test is conducted, the experimental data is used to validate the
model results, within a certain thermal uncertainty margin. NASA GEVS requires a
10◦ C margin for qualification/protoflight testing [1] and The European Cooperation
for Space Standardization (ECSS) requires a 15◦ C [1, 11] margin for the same testing.
These margins are for thermal test levels above or below the predicted flight levels,
so they are not exactly analogous to comparing the thermal model of the chamber
to the response of the chamber. However, they show that the worst case prediction
should have an error less than 10-15◦ C.
9
1.5 Experimental Apparatus Background and Previous Capabilities
The MDA US systems thermal vacuum chamber, referred to as the “Blue” TVAC was
donated by MDA Corporation and was originally used in the development of the Mars
exploration rover’s robotic arm [11]. Since it was donated, significant refurbishment,
analysis, and testing has been performed by multiple graduate and undergraduate
students to improve the functionality and capabilities of the system. Figure 1.4 from
Jensma’s thesis shows the Blue TVAC in its state before any work in this thesis was
performed [11]. When the previous state is referenced, this refers to the state of the
chamber before any modifications for this thesis.
Figure 1.4: Blue TVAC Chamber [11]
The Blue TVAC consists of three primary systems: the vacuum system, the ther-
mal control system, and the data collection and control systems. The vacuum system
10
consists of pressure gauges and two pumps: one turbomolecular pump, and one me-
chanical pump. The thermal control system consists of a platen with heaters and
coolant lines, a shroud with coolant lines, a door shroud, and nitrogen manage-
ment systems including valves and a dewar. The data collection system includes
feedthroughs for thermocouples, a temperature controller to manage heaters and ni-
trogen flow, and a DAQ system with 12 thermocouple channels. A system level block
diagram from Jenmsa’s thesis is shown in Figure 1.5 [11].
Figure 1.5: Blue TVAC systems block diagram [11]
The previous capabilities of the Blue TVAC as of June 2020 are adequate for up
to two full thermal vacuum cycles with the mass model according to ISO Standard
19683 at a minimum vacuum pressure of <2.0e−4 Torr [11]. The previous minimum
vacuum pressure was not compliant with ISO Standard 19683 for thermal vacuum
testing but was assumed adequate for the assumption that convective heat transfer
is negligible [11]. The previous lowest and highest achievable temperatures during
thermal vacuum testing were approximately -30◦ C and 50◦ C for the shroud, and -
20◦ C to 65◦ C for the platen [11]. The existing DAQ has 12 thermocouple inputs
11
for measuring shroud, platen, and test article temperatures throughout a test along
with data logging capability [11]. Table 1.1 summarizes the Blue TVAC’s previous
capabilities compared to ISO 19683. Note that tolerance limit is the minimum and
maximum difference of the test article from the set point during thermal dwells. It is
difficult to precisely compare this value between different experimental setups due to
a difference in where temperatures are recorded and the limitations of only logging
temperature at certain locations along with varying definitions of steady state.
ISO 19683 Previous Description

Require- Capability
ment [9] [11]
Temperature -15◦ C to -15◦ C to For cold soaks, only capable
◦ ◦
Range +50 C +50 C of a couple cycles due to liq-
uid nitrogen tank size and
consumption rate [11].
Chamber <7.5e−6 Torr <2.0e−4 Torr Further leak investigations
Pressure could improve this.
Number of 2 or More 4 Greater is better, modifica-
Cycles tions should increase this.
Thermal 1 hour or 1 hour or Modifications should in-
Dwell longer longer crease this.
Temperature ±5◦ C or ±5◦ C or Greater is better, modifica-
Ramp Rate slower slower tions should increase this.
Tolerance 3◦ C Not Specified Modifications should allow
Limit this to be met.
Table 1.1: Previous Capability vs ISO 19683 Requirements for TVAC

Testing
Table 1.2 shows a further breakdown of the various levels, margins, and tolerances
for thermal vacuum testing from various rules and standards. Note that the aver-
age tolerance is +/-3◦ C and a chamber pressure requirement is <10−4 Torr for all
standards shown.
The thermal control system consists of the shroud, platen, and door shroud. All
these components are plumbed such that nitrogen can flow through them for cooling
and the flow can be controlled using a PID controller and solenoid actuated valves.
12
The platen has electric resistance heaters mounted underneath it that are controlled
using a PID controller. The shroud is made up of copper sheeting with soldered on
copper tubing and is uncoated. The platen has built in coolant channels and heaters
installed underneath and is made of uncoated aluminum. The door shroud is likely
aluminum with coolant channels and is coated with a flat black coating on the inside
face. Liquid nitrogen is fed from an external 35 L Dewar and the outlets from the
chamber through tubing to the atmosphere outside of the lab.
1.6 Limitations and Modifications
The major limitations with the Blue TVAC in its previous state were maximum po-
tential test article size, minimum pressures, maximum and minimum temperature
ranges, and cold hold duration. These were persistent limitations with the chamber
since it was donated and as it was upgraded, and extensive testing and characteriza-
tion of the chamber has been performed by previous students to model and improve
the performance and consistency of the chamber. Caudill and Diamond implemented
the cooling system, and they demonstrated the shroud cools more rapidly than the
platen, likely due to thermal mass and plumbing differences [3]. It was also noted
that times to reach thermal equilibrium for a 3U CubeSat mass simulator were on the
order of hours [3]. This is likely due to poor heat transfer between the shroud, platen,
and test article. The shroud is bare copper, which has an emissivity of less than 0.1
depending on the surface roughness and presence of oxidation [11]. The platen is bare
aluminum, which also has an emissivity of typically less than 0.1 [11]. This reduces
the rate of radiative heat transfer to the test article compared to a higher emissivity
surface, as the shroud and platen emissivity is directly proportional to the amount of
heat it can absorb or emit.
13
Both maximum and minimum temperature capabilities and ramp rates can be
improved by decreasing the shroud and platen thermal mass and adding high emis-
sivity coatings to the interior faces of the shroud and platen. High emissivity coatings
in this thesis are regarded as those with an emissivity of greater than 0.8, which is
typical of aerospace grade black coatings [8]. Research by Jayaram and Gonzalez has
demonstrated a low cost, small thermal vacuum chamber can be constructed using
common materials with comparable performance to existing commercially available
chambers [10]. They utilize copper sheeting and copper refrigeration tubing to con-
struct a cylindrical shroud, with a small platform to accommodate test articles. This
is similar to the previous configuration, except with a lower thermal mass relative to
the size of the chamber and a larger maximum test article size relative the pressure
vessel size. Based on the results from Jayaram and Gonzalez, the reduction in thermal
mass of the Blue TVAC shroud and platen will offer performance improvements due
to the lower thermal mass, which will reduce time to equilibrium temperatures and
use less nitrogen to achieve a given set temperature. The shroud plumbing will need
to be resized with thinner materials and tubing to achieve the lowest temperatures
possible with a backpressure of 27,579-48,263 Pa (4-7 psi) [11, 3], 35 L storage capac-
ity, and evaporation rate of 0.22 L/day [10]. Additionally, improving the insulation
around the dewar feed lines to reduce thermal losses based on the losses noted by
Jensma, Caudill, and Diamond during testing. The existing platen heaters must be
repurposed as heater plates such that they can be arranged around the test article
as needed to simulate an IR source like the Earth or other planetary bodies, or an
approximation of the sun’s heat input. This can be achieved by setting the heaters
such that they provide the equivalent heat flux that would be seen in flight from the
sun [1].
In addition to reducing the thermal mass of the heating/cooling system, coatings
can be utilized to improve the performance and efficiency of the system. By applying
14
high emissivity coatings to the interior faces of the shroud and heater plate surfaces,
equilibrium temperatures for the test article are expected to be achieved in less time,
and as a result less heater power and nitrogen is needed to achieve a given test article
temperature. Extensive testing of thermal control materials has been conducted at
the International Space Station on the Long Duration Exposure Facility to evaluate
the performance of commonly used coatings for thermal control. Specifically, low
outgassing polyurethane thermal control coatings like Aeroglaze Z306 were resistant
to the space environment and even increased in emissivity likely due to roughening by
erosion from Micrometeorites and orbital debris (MMOD) and atomic oxygen (AO)
[10]. While the TVAC chamber will be free of AO and MMOD, this polyurethane
coating will likely be able to survive the extreme temperatures, thermal cycling, and
vacuum environment in the chamber with minimal degradation and outgassing/con-
tamination risk.
Since the likely leak noted previously is persistent and increasing, it is important
that the leak is investigated thoroughly, and resolutions are made as appropriate.
The space environments lab is equipped with a helium leak detector, which was used
to attempt to locate the major leak sources as part of the work performed for thesis.
It was not documented by Williams or Jensma that the helium leak detector has
been used with the Blue TVAC before to help determine the root cause, and Jensma
recommended its use to find leaks in the system [22, 11].
1.7 Thesis Motivation and Work
After hardware changes and leak testing discussed in the previous section were made
to the Blue TVAC, TVAC testing and thermal modeling was needed to confirm the
changes made improved performance and capabilities. A new thermal model using
SolidWorks and SolidWorks Flow Simulation was updated to match the current con-
15
figuration and compared to testing results such that a predictive analytical model
can be made. This testing involved measuring the temperature response of different
components of the system as well as test articles during thermal vacuum testing.
Test articles include a CubeSat mass model and a coated and bare aluminum plate,
which are geometrically simple and well characterized in terms of surface and material
properties.
The goal and motivation of this thesis is to improve the performance of the system
and re-calibrate the models and control systems to the new configuration such that it
meets the ISO 19683 test level requirements for thermal vacuum testing. This work
builds off the efforts of previous students so that the chamber can be used as an
educational and research tool for projects within and outside of Cal Poly. In general,
the chamber in its current state meets ISO 19683 requirements except for vacuum level
and tolerance. However, the capabilities can be extended with the modifications which
should increase test volume, reduce the leak rate and therefore minimize chamber
pressure, improve thermal mass such that the nitrogen consumption rate is decreased
which enables longer dwells and more cycles, and improve the radiation efficiency
of the heaters and shroud with coatings which will enable and faster equilibrium
times and more efficient heat transfer to test articles. These improvements, while
not strictly based on any requirements beyond ISO 19863, are important because
they offer more flexible test capabilities. These more flexible capabilities are able to
accommodate full scale, extended tests in the case of larger CubeSats, for example.
16
Table 1.2: Rules and standards for thermal vacuum cycling tests from
Chisabas et al. [4]
17
Chapter 2
MODIFICATION DESIGN DEVELOPMENT
The changes to the Blue TVAC heating and cooling system were motivated by the
efforts of previous students and build off their work and incorporate the future work
specified in previous theses by Williams and Jensma [22, 11]. The focus of this chapter
is on sizing and optimizing the cooling shroud, as well as designing for modularity
and manufacturability using vacuum compatible materials.
2.1 Goals and Design Criteria
In order to improve the performance of the heating and cooling systems, the entire
existing platen and shroud assemblies required design from the ground up. The
modified design was developed with funding and manufacturing limitations in mind.
In general, parts are reused and modified where possible, and manufacturing is limited
by student capabilities and the tools available in the Aerospace Engineering shop and
other shared work spaces on campus. The goals of the modifications are to:
• Expand the available space for test articles by incorporating a larger shroud
and modular components
• Improve the temperature distribution across the shroud and platen such that
test articles are evenly heated/cooled
• Reduce the thermal mass of the chamber to improve response time of the cham-
ber and require less energy to heat and cool by using thinner materials where
appropriate, allowing for greater numbers of thermal cycles and/or greater tem-
perature range per unit of liquid nitrogen consumed
18
• Improve the heat transfer rate between the test article and the chamber systems
through the use of high emissivity coatings
• Improve the minimum pressure achieved using vacuum compatible materials
and a leak tight cooling system
• Allow the thermal response to be more predictable through the creation of a
validated thermal model using SolidWorks
The modification goals would be considered successful if there was a noted im-
provement in all the mentioned criteria above while also maintaining and exceeding
ISO 19683 requirements for thermal vacuum testing.
2.2 Cooling Shroud Sizing
It was determined that a cylindrical cooling shroud would make the most efficient use
of space in the chamber because the existing pressure vessel is cylindrical in shape.
Additionally, fabricating a cylinder shape allows for simple tooling compared to other
geometries. After an initial shape was determined, a detailed study was performed
to determine how to direct the nitrogen coolant throughout the shroud such that it
allows for an even temperature distribution across the interior surface.
2.2.1 Initial Configuration Comparisons
It was determined that using commercially available copper tubing, common for re-
frigeration, HVAC, and water distribution applications would be best suited for this
application. This tubing is generally high purity copper, which is desirable for out-
gassing properties, corrosion resistance, thermal conductivity, and workability. After
a base material was selected, different configurations were conceptualized and mod-
19
eled using SolidWorks. Figure 2.1 shows the four main options considered for the
tubing configuration.
Figure 2.1: Concept shroud cooling tubing configurations
The single spiral configuration consists of continuous tubing from one end of the
shroud cylinder to the other, wrapping around the outside of a thin sheet of copper.
This is the simplest configuration because there is only one inlet and outlet, one
continuous shape, and commercial tubing can be used directly with little modification.
The double spiral consists of interwoven spirals, where one outlet and one inlet exists
on each end of the shroud cylinder. The concept behind the double spiral was that a
20
more even temperature distribution could be achieved with inlets and outlets on both
sides, but at the cost of using the coolant less efficiently since the warmer nitrogen
exiting the shroud would be in close proximity to the incoming nitrogen from the
other spiral segment. The flattened tubing configuration is a single spiral without
any underlying sheet of copper to act as the shroud cylinder. Large tubing would
be flattened such that the tubing is wide along the axis of the cylinder, and thin
in the axis normal to the cylinder. This configuration requires tubing that can be
easily formed into this shape without damaging the tubing such that it would leak or
fail to hold pressure. The flattened tubing configuration is shown as a double spiral,
but could be a single spiral as well. Finally, the manifolds with rings configuration
consists of an inlet manifold on the top of the cylinder and an outlet manifold at the
bottom of the cylinder. These manifolds run along the walls of the cylinder and carry
nitrogen coolant to the semi-circumferential tubing that runs from the inlet to the
outlet manifolds along the walls of the shroud.
Further research into available materials lead to the elimination of the flattened
tubing configuration. This is because the available tubing sizes that would be able
to be flattened as shown in Figure 2.1 have wall thicknesses that would result in
excessive mass compared to other configurations, which is undesirable because the
thermal mass of the cooling system would significantly increase.
The manifolds with rings configuration requires many tens of joints that must be
entirely leak tight, or the chamber will have nitrogen leaking into it, which the pumps
would likely not be able to keep up with. Additionally, the number of precise holes
in the manifold and subsequent joints means the configuration is more complex in
terms of manufacturability than the spiral and double spiral configurations. Because
of this, the manifolds with rings configuration was eliminated as well.
21
2.2.2 Detailed Configuration Comparison
After the configurations were narrowed down to the single and double spiral con-
figurations,study was performed using SolidWorks Flow simulation to compare and
optimize the spiral configurations, focusing on even temperature distributions across
the shroud while seeking to minimize material used and manufacturing complexity.
2.2.2.1 Numerical Model Setup
SolidWorks Flow Simulation was used for the configuration comparisons and consisted
of a simplified chamber geometry to simulate heat moving from the environment to the
chamber as well as the shroud and test article to examine temperature distributions
across the various surfaces in the cold, steady state case.
An 18” outer diameter 304 stainless steel cylinder was used to approximate the
vacuum chamber, with a polished interior face. 0.75” of fiberglass insulation around
the chamber cylinder section was included to approximate the insulation installed
on the chamber. All exposed chamber and insulation surfaces were assumed to be
at a constant temperature boundary condition of 20◦ C. This is an approximation of
the average temperature in the lab based on day/night cycles where the chamber is
located. Because there are multiple buffer zones between the outside and the test
article, like the building walls, air gap, chamber insulation, chamber walls, vacuum
gap between the shroud and chamber, external temperature variation is not expected
to play a significant role in measured thermal response. Additionally, the chamber
is actively controlled which will further damp variation in experimental results. The
emissivity used for polished stainless steel was 0.074 and is from the SolidWorks
standard material library for polished stainless steel [20].
22
The coolant used is nitrogen and is assumed to be gaseous at 100 K when it
enters the tubing (77 K boiling point + arbitrary losses). The nitrogen flow rate
was estimated to be 49 g/min [11]. This is based on estimates of the duty cycle
and endurance of the 35 L dewar used in Jensma’s testing. The backpressure for the
cooling system outlet is assumed to be 0.6 m of water, which is based on previous
test configuration [11].
The copper sheet used to support the tubing and absorb energy from the test
articles was defined as 0.005” copper sheeting with a diameter of 16.5” and a length
of 30”. The shroud tubing consisted of standard 1/4” OD copper refrigeration tubing.
A tubing spacing of 2” for the spirals was selected for the initial single and double
spiral comparison. The sheet and tubing are bonded with no contact resistance and
the interface completely connects the thickness of the shroud to the base of the tube.
The emissivity used for polished copper was 0.025 at 300 K and is a function of
temperature. The emissivity is from the SolidWorks standard material library for
polished copper [20].
The test article simulator used was a simplified 3U CubeSat mass model made
of black painted aluminum with 1/8” wall thickness and a heat dissipation of 5 W
throughout its volume. An emissivty of 0.91 was assumed for the coated surfaces
including the inside walls of the shroud and CubeSat.
There is no contact between the CubeSat and shroud or shroud and chamber
interior to simplify losses through supports so that radiation is the only means of
heat transfer between the surfaces.
2.2.2.2 Single and Double Spiral Comparison
Because this model compares temperature distributions and the only parameter being
changed is the tubing configurations, direct comparisons can be made between the
23
results at steady state. The actual temperatures in this simulation are not intended to
be compared to the experimental results since the coolant flow rate was approximated,
but the intent is that the temperatures will be representative and comparable between
different configurations. In Chapter 4 and 5, the thermal model used for experimental
results comparison is discussed. Because the temperature distributions do not vary
by more than 1◦ C across the surface, contour plot results are not shown.
Figure 2.2 shows the temperature distribution across the shroud for the double
spiral configuration. Note that the middle region has the smallest difference in tem-
peratures, around -144◦ C to -150◦ C, and the edges farthest from the tubing tend
to have the greatest difference in temperatures compared to the average, around -
136◦ C to -156◦ C. The maximum temperature difference across the shroud is 29◦ C,
but the majority of the shroud is between -145◦ C to -155◦ C. The test article surface
temperature is maintained to within 0.13◦ C.
Figure 2.3 shows the temperature distribution across the shroud for the single
spiral configuration. Note that the middle region has the smallest difference in tem-
peratures, around -180◦ C to -184◦ C, and the edges farthest from the tubing tend
to have the greatest difference in temperatures compared to the average, around -
162◦ C to -175◦ C. The maximum temperature difference across the shroud is 22◦ C,
but the majority of the shroud is between -180◦ C to -184◦ C. The test article surface
temperature is maintained to within 0.13◦ C.
It is clear that the minimum temperature for the majority of the shroud is lower
for the single spiral, and this because the double spiral is less efficient at cooling
because of the proximity of different spiral tubing, where the inlet side of one spiral is
close to the outlet side of the other spiral. There was no improvement in test article
surface temperature difference between the single and double spiral cases, and the
minimum temperature was lower for the single spiral. The single spiral also had less
24
Figure 2.2: Double spiral temperature contour plot for the shroud with
2” spacing
25
Figure 2.3: Single spiral temperature contour plot for the shroud with 2”
spacing
26
variation in temperature across it, which is clear from the contour plots. Overall,
the single spiral offers similar or better performance with less complexity, and will be
further optimized in the next section.
2.2.3 Single Spiral Optimization
Following a similar methodology for model setup from the previous section, the single
spiral configuration was compared with various changing parameters including the
spiral spacing, the tube size, and the shroud sheet thickness. First, the tubing spacing
was compared at 1”, 2”, and 3” between spirals to determine the sensitivity of spacing
on temperature distribution and minimum temperatures. Figures 2.4 and 2.5 show
the temperature distributions at the 1” and 3” spacing, with 2” being the same as
2.3 from the previous section.
The different configurations shows smaller spacing leads to more consistent tem-
perature distributions, with a minimum temperature of -190◦ C at 1” compared to
-184◦ C at 2” and 3”. All cases showed test article surface temperature is maintained
to within 0.14◦ C. Because the single spiral requires over 100 feet of tubing and has a
negligible affect on test article surface temperatures in this case, it was determined
that a spacing between 2” and 3” would be selected.
In addition to tubing spacing, shroud thickness and tubing size was varied. Figure
2.6 shows the effect of increasing the tubing OD from 1/4” to 3/8”. The minimum
and maximum range increases and the minimum temperature is higher compared to
the 1/4” tubing case, so the 3/8” tubing was not used. Sizes smaller than 1/4” OD
were not studied in detail because there are not standard, commonly available sizes
below 1/4”.
The effect of thicker shroud sheet was studied by increasing the shroud sheet
from 0.005” to 0.008” thickness. Figure 2.7 shows the effect on temperatures across
27
spacing
28
spacing
29
spacing and 3/8” OD tubing
30
the shroud and platen. The effect is insignificant because the minimum temperature
changes by 0.19◦ C and the maximum and minimum temperatures are less than 2◦ C
different. Because of this, the 0.005” shroud sheet was selected due the lower thermal
mass and easier formability.
spacing and 0.008” thick sheet
2.2.4 Configuration Selection
Shroud configuration was selected based on comparisons and rankings of different
factors shown in Table 2.1. The shroud maximum temperature difference is the
31
difference in maximum and minimum temperatures across the entire shroud surface.
Temperature distribution rank is a ranking from 1 to 6, where 1 is the best, of
how even the temperature is across the majority of the shroud surface. Minimum
temperature rank is the lowest achieved temperature for a given configuration, where
1 is the best or lowest temperature. Manufacturability rank considers how difficult
it is to manufacture a given configuration, considering length of tubing, geometrical
complexity of configuration, and material thickness.
Table 2.1: Cooling sizing analysis summary table, best configurations high-
lighted in green
The final configuration was selected with 1/4” OD tubing and a 0.005” shroud
sheet thickness was selected based on the above results. The spacing ultimately used
for manufacturing was 2.75”, because the tubing came in a roll at a fixed length
and 2.75” allowed for it to be fully consumed without increasing the copper tubing
length required to be purchased. The final length of the shroud cylinder was 24” to
allow for 4” at the front of the chamber for the thermocouple input board as well as
plumbing connections for nitrogen. The mass of this modified shroud is less than 3
kg, compared to the previous shroud with a measured mass of over 11 kg.
32
2.3 Heater Plate Sizing
The heater plates were sized to minimize thermal mass while still maintaining rigidity
and allowing for an even temperature distribution across the plate under expected
conditions with the heater element operating. It was determined that three heater
plates were to be cut from the existing platen, and were sized to be capable of sur-
rounding a 3U CubeSat on three of the largest faces. The existing platen material
was used to save cost and minimize the risk of damage to the strip heaters integrated
in the platen, as they were press fit into slots in the platen.
The final dimensions for the heater plates were 6” by 12” with a 0.14” wall thick-
ness and 0.25” wide slots to mount the heater strips. The heating elements are 10” x
0.25” x 1.5” and cover over 20% of the back surface area. The mass of the modified
heater plates was approximately 0.5 kg each or 1.5 kg total, compared to the previous
platen which weighed in excess of 20 kg. These are not directly comparable since the
platen also provided a cooling function that is replaced by the cylindrical shroud, but
give an idea of the difference in thermal mass of the system.
2.3.1 Heater Plate Thermal Model
To verify the temperature distribution across the front face of the plates would be
even with these dimensions, a simple thermal analysis using SolidWorks simulation
was performed. The analysis assumes the heater strip acts as an 85◦ C constant
temperature boundary condition for the slot where the strip is pressed into. The
faces except for the slot for the heater strip is painted black with an emissivity of
0.91 and is assumed to have a view factor to the ambient environment of 1 with an
environment temperature of 65◦ C. These conditions are intended to represent a worst
case where the test article and surrounding shroud are 20◦ C colder than the heater
33
strip temperature. The resulting front face temperature distribution contour plot is
shown in Figure 2.8. Because the surface temperature varied by less than 1◦ C, the
modified heater plates approximate a constant radiating surface temperature that a
test article will be exposed to.
Figure 2.8: Contour plot of heater plate front face temperature distribu-
tion
It was determined that placing the heater plates at the bottom of the chamber as
shown in Figure 2.9 allows for an even temperature distribution in the chamber and
keeps the modular platen free to be used exclusively for test articles. Placing them
at the bottom of the chamber means that they are able to radiate to a test article
from underneath as well as to the sides and bottom of the shroud, which conducts
the heat around the shroud and allows for an even temperature distribution across
the shroud. This configuration can be modified since the heater plates are modular
and not permanently mounted.
2.4 High Emissivity Coatings
For this thesis, one gallon of Aeroglaze Z306 was donated by Socomore and other
supporting materials including primer and thinner were also purchased or donated
34
from Socomore. Aeroglaze Z306 has extensive flight heritage and has been extensively
studied and used for spaceflight applications [8]. The coating has a typical emissivity
of 0.91, an ASTM E 595-77 TML of 1.0% and CVCM of 0.02%, and an operating
temperature range of -150◦ C to 130◦ C [19]. The high emissivity improves heat transfer
via radiation between the shroud, heater plates, and test articles to allow for more
even temperatures and faster equilibrium time.
2.5 Platen, Modular Mounting System
The platen selected based on commercially available oven racks or cooling racks.
Oven racks are typically constructed from polished stainless steel, which is vacuum
compatible. For this thesis, two 10”x15” oven cooling racks were selected and modified
to remove unnecessary parts. The open area of the platen wire mesh is approximately
71%, and is made up primarily of 1/16” 302 stainless steel wires in a square pattern,
with larger wires around the perimeter and in the middle for support. The open area,
in combination with the reflective surface, means that heat can transfer via radiation
between the bottom of the chamber to the test article on top of the platen. The
spacing between the mesh is approximately 3/8” with wire diameters of approximately
1/16th of an inch. Mounting bolts with washers are recommended to secure test
articles to the platen.
The modular mounting system consists of modified, commercially available alu-
minum rails. The rails selected are modified low profile T slotted framing rails. They
have a width of 1” and depth of 0.5” and span the entire 24” length of the chamber
with 4 rails spaced 90◦ apart. These allow for platen mounting as well as flexibility for
test article and heater plate mounting depending on the requirements of a given test.
For example, the platen could be removed entirely and a test article could be hung
from the top mounting rail using a custom bracket or brackets to allow for larger test
35
article. The aluminum rails can accommodate any nuts compatible with 1” 80/20
t-slotted framing.
2.6 Aluminum Foil Reflectors
High vacuum compatible aluminum foil was utilized in the back of the shroud, over
the door shroud, and at the front of the shroud to act as a radiative barrier between
the test article, shroud interior, and the outer chamber. This saves energy or nitrogen
and allows for more even temperature distributions on the shroud and test articles.
The reflectors are shown in the final modification design section.
2.7 Final Modification Design
The final modification design is shown in Figure 2.9. It incorporates the findings from
the cooling system sizing and includes a 3U CubeSat mass model test article for scale.
Figure 2.9: Final modification design section view with labeled compo-
nents and a CubeSat mass model test article shown
36
2.8 Liquid Nitrogen Tank Upgrade
For this thesis, an R-IDC grant was secured by Dr. Abercromby for the purpose
of upgrading the existing 35 L liquid nitrogen dewar, which cost $1,925 with freight
shipping. The funding allowed for the purchase of a 180 L liquid nitrogen tank and
supporting hardware, including a scale and a dolly. This represents a stored liquid
nitrogen capacity increase of over 400% which offers greater flexibility and capabilities
for extended duration testing, including lower temperature dwells, extended dwell
periods, and/or greater thermal cycles.
The tank is a Dura-Cyl LC180-22, and comes with ports, gauges, and safety
features required for operation. The existing cryogenic liquid hose used with the 35
L dewar can be directly connected to the liquid withdraw outlet on the 180 L dewar
and feed into the chamber plumbing. The 180 L tank comes with a relief valve and a
burst disk to prevent over-pressure. The tank automatically maintains a pressure of
approximately 21-22 psi by venting gaseous nitrogen through the relief valve as the
liquid slowly evaporates. The tank has an evaporation rate of 1.5% which corresponds
to a loss of approximately 2.7 L/day. This means that liquid nitrogen can be stored
without significant losses for days while it is in use. The cost to refill the tank from
Praxair was approximately $86 as of June 2021. There is 1 week lead time on the
refills from Praxair, as they pick up the tank on Thursdays and return it filled the
following Thursday. It is possible other gas suppliers like Airgas could be used if
shorter lead time is required, but this was not investigated in detail.
37
Chapter 3
MANUFACTURING PROCESS
As mentioned in the previous section, manufacturability played a significant role in
the design of the updated cooling and heating systems, as well as re-usability of
existing parts and materials due to budget limitations. Despite this, margin in both
manufacturing timelines and budgets should always be included and were needed
for this work. The greatest challenge in the manufacturing process was the shroud
assembly, specifically the brazing process, which will be outlined along with other
major parts of the manufacturing process.
3.1 Shroud Assembly
3.1.1 Copper Tubing and Sheet Properties
The copper tubing was purchased in coil form from McMaster Carr, and is typically
used in refrigeration, water, and oil and gas industries [14]. The tubing is soft 122
copper with a purity of 99.9% with trace amounts of phosphorous to displace oxy-
gen [14]. Outer diameter is 1/4 inch with a 0.003” wall thickness, and is rated for
1,400 psi at room temperature [14], and given the maximum operating pressure of
approximately 14.7 psi, no additional analysis was performed to verify it would hold
pressure, as copper becomes stronger and more ductile at lower temperatures [5].
The copper sheet was made from A110 copper, which has 99.9% purity, and was
purchased from basiccopper.com in a 24” x 6’ x 0.005” roll [2]. It could be formed
by hand to the previous dimensions and cut using metal shears or a foot shear due
to its thickness.
38
3.1.2 Tubing Forming
It was desirable to increase the contact area with the shroud copper sheet to improve
heat transfer from the nitrogen flowing through the tubing to the copper tubing. To
accomplish this, an electric roller was used to deform the tubing to have two flat faces
and constant radius, semicircular edges, similar to a slot. The final dimensions for
the formed tubing are a height of 0.17” with a flat width of 0.15” and a semicircular
radius of 0.08”. A labeled section of sample tubing is shown in Figure 3.1.
Figure 3.1: Tubing cross section after rolling process
After the tubing was rolled to the desired cross section, a ring roller was used to
roughly form the tubing such that it has a 16.5” inner diameter and is spiral in shape
so that it can be brazed to the outside of the copper sheet. Final forming takes place
when the tubing is tacked and brazed to the shroud sheet. Since the tubing is soft
copper, it can be formed by hand to meet final required radius of 16.5” and spacing
of 2.75”.
39
3.1.3 Shroud Assembly Brazing
The brazing of the shroud tubing to the copper sheet was the most involved and time
consuming part of the manufacturing process. This process required the brazing of
over 40 feet of tubing in a single spiral around the 16.5” diameter, 24” long sheet.
The brazing alloy used for this process was an alloy of 45% silver, 30% copper, and
25% zinc in strip form 1/2” wide by 0.005” thick. An oxyacetylene torch was used to
heat the tubing and sheet for brazing, and appropriate white brazing flux was used to
prevent oxides from forming and inhibiting the connection during the brazing process.
Ideally, the shroud surfaces to be brazed would be cleaned and abraded with
Scotchbrite or similar abrasion, however, it was found with coupon testing that for
this application cleaning was not required as the raw materials were clean enough that
flux alone allowed for an adequate bond. Bonds were periodically checked during the
brazing process by visual inspection and physically pulling on the joint to ensure it
does not separate by hand forces on the order of 10-20 lbf. This is required because
when the flux melts, it forms a dark, glassy substance that can obscure the joint
visually and physically joins the brazed parts, so the parts can appear joined when
in reality only flux holds them together. An adequate bond has a consistent, visible,
and polished looking fillet indicating that the brazing material has joined the two
parts cleanly and evenly.
To become familiar with the brazing process using these specific materials, test
coupons were created first using scrap copper sheet and tubing, then using cutoffs of
actual materials used for the project, to verify the methods used will not damage the
final materials and to practice brazing technique. It was noted that during brazing,
heat should be focused on the tubing and not the sheet. This is because the tubing is
more massive and takes heat away from the brazing area. Moreover, directly heating
the sheet can cause overheating and melt the copper, burning a hole through the sheet.
40
It was determined that a copper sheet patch could be brazed on to burned through
areas if repairs were needed. It was also noted that brazing circular tubing to the
sheet required more brazing material vs a flattened tube with a smaller radius fillet.
Figures 3.2 and 3.3 show the brazing coupons. In Figure 3.2, the joint is consistent,
even, and reflective but is relatively large because the tubing is not flattened, while
in Figure 3.3 the bottom quarter of brazed joint has a relatively smaller fillet because
of the flattened tubing. The fillet is not visible past the bottom quarter and that
section would be considered inadequate since a fillet is not fully visible.
Figure 3.2: 1st brazing coupon with round tubing and thicker copper sheet
Note that in the 2nd brazing coupon, the fillet is not fully formed in some locations
due to the technique used. Later during brazing, the technique changed, as discussed
later, allowing for the fillet to be maintained along the majority of the interface.
To begin final brazing, the copper sheet was trimmed to the desired circumference
such that the diameter is 16.5 inches and overlapped by approximately 1/4” and
clamped on its ends, then brazed together from the center outwards. Figure 3.4
shows the sheet during seam brazing. Note that because of the thin nature of the
sheet, it does not hold its shape without the supporting copper tubing.
41
Figure 3.3: 2nd brazing coupon with flattened tubing and 0.005” copper
sheet
Figure 3.4: Copper sheet during brazing process for cylinder forming
42
After the sheet was joined, the diameter was checked, and the formed tubing was
roughly fit onto the sheet. The tubing provides reinforcement and allows the rough
shape of the shroud to be seen. This is illustrated in Figure 3.5.
Figure 3.5: Copper coil roughly fit onto brazed sheet
After the initial connection of the copper tubing to the shroud sheet was made,
different methods of brazing the sheet and tubing were compared so that the brazing
process would use the least amount of brazing material possible and provide a seamless
connection between the tube and shroud. The best method in this case involved tack
brazing the coil every 1/4 to 1/3 rotation using a spacer gauge to maintain the 2.75”
tube spacing. After tacking, the space between tacks was filled by placing 0.005” thick
brazing rod sheet between the coil and sheet, and heating the joint until the brazing
material liquefied. Then, using a ball peen hammer on the inside sheet and the torch
on the outside to keep the joint liquid, the joint was hammered together along the
space between tacks. Because the shroud sheet is thin and malleable, it forms around
the bottom of the tube and squeezes out excess brazing material, creating a thin joint
and allowing for consistent contact along the length of the interface. Figure 3.6 shows
the state of the shroud about halfway through the brazing process.
43
Figure 3.6: Shroud during brazing process
Note that the brazing process causes oxidation of the shroud interior and can
be seen by the darkening of the chamber interior. Additionally, the outside of the
chamber shows oxidation and flux residue on and near the tubes length that has
been brazed. The tubing was purged with constant flow of gaseous nitrogen during
brazing to prevent oxide formation inside of the tubing, which would be difficult to
remove. After brazing, a thorough cleaning method was needed to remove brazing
fluxes and oxides on the exterior faces. These had to be removed because brazing
fluxes are known to outgas significantly, oxides are known to absorb water vapor,
and the impact of contamination on primer and paint adhesion was unknown [13].
A solution of hot dilute citric acid and water was used to remove brazing flux and
oxidation. Citric acid is the safest acid recommended for use by the flux manufacturer
for thorough flux removal. The citric acid solution consisted of approximately 1 lb of
citric acid and 6 gallons of water. The shroud was placed in the solution and scrubbed
44
with a stainless steel wire brush. Figures 3.7 and 3.8 show the shroud during cleaning,
and after cleaning, respectively.
Figure 3.7: Shroud in plastic tub during cleaning with dilute hot citric
acid
Before shroud painting, the modular mounting rails used to support the heaters,
platen, and test articles were fit and mounting holes in the shroud were made. The
rails are attached to the shroud using six stainless steel 10-32 machine screws and
stainless steel washers. Each rail position was measured, then holes were made in
the shroud and rails for the bolts. Each bolt was test fit and marked to be trimmed
so that it did not protrude into the interior slot of the rail, where it could interfere
with anything being mounted to the rails through the slots. Figure 3.9 shows one rail
with protruding bolts in the slot. Before the rails were test fit, the anodize coating
was removed using oven cleaner, which contains sodium hydroxide. This method was
used because it is common practice, utilizes commonly available cleaning products,
and is safe to use with proper personal protective equipment [16]. The coating was
45
Figure 3.8: Shroud after cleaning process with dilute hot citric acid
removed to reduce outgassing of adsorbed water vapor, and the profile was machined
to reduce the thermal mass while maintaining functionality.
3.1.4 Shroud Assembly Painting Coupon Testing
After citric acid cleaning, the shroud was ready for paint preparation. The nominal
application method utilizes thin coats of primer and paint using an HVLP (high ve-
locity, low pressure) paint sprayer to apply thin coats of primer and paint [19, 18].
Coupon tests were performed to determine the best method to apply the Aeroglaze
9947 wash primer and Aeroglaze Z306 flat black polyurethane coating. During coupon
tests of primer application using an HVLP paint sprayer, significant fisheyes or pin-
holes were visible on the surface, causing uneven application of primer to the surface.
This are seen in Figure 3.10. In addition to the fisheyes and pinholes, the primer
was applied thicker than specified, since the substrate should be visible through the
primer, and should only leave a slight orange tinge to the surface. The opacity of
46
Figure 3.9: One mounting rail test fit and bolted in place to determine
bolt lengths
the coating indicates it is applied thicker than specified, which could lead to poor
adhesion of the topcoat.
In order to determine the source of the contamination, the manufacturer was
consulted and provided recommendations for cleaning methods and possible contam-
ination sources. Initially, when applying primer, acetone and IPA was used to clean
the surface in preparation for painting. The manufacturer recommended performing
a water break test after cleaning to approximate how clean the surface is. Using
copper sheet coupons, one coupon was cleaned thoroughly using acetone, IPA, then
using HET Grease, Wax, and Silicone Remover, then a final clean with IPA, and the
other coupon was not cleaned for control comparison. Figure 3.11 shows the water
break test before and after cleaning. The surface after cleaning was not fully water
break free despite the thorough cleaning.
47
Figure 3.10: First primer coupon primer application with fisheyes and
excessive coating thickness
Figure 3.11: Copper sheet coupon water break test before cleaning (upper)
and after cleaning (lower)
48
There are a variety of potential causes of this residual contamination indicated by
the water break test, including but not limited to:
• Residual, hard to remove silicone contamination embedded in copper sheet sur-
face from unknown factory contamination source
• Contamination in one of the cleaning solvents, like residual silicone or oil
• Contamination from environment around cleaning area (cleaning and painting
took place in a machine shop and outdoors)
Additionally, contamination from residual paint in the paint sprayer itself as well
as oil or other contaminants in the air lines were possible sources during spraying.
Cleaning issues were not further investigated because after the thorough cleaning
method above and wiping off excess primer after spray application, an acceptable
surface quality was achieved. For shroud painting an additional final cleaning using
Aeroglaze 9958 paint thinner was performed after the 2nd IPA cleaning step. A
second primer coupon was prepared using the new cleaning method as well as wiping
off excess primer, and the results are shown in Figure 3.12
The 2nd primer coupon was then painted with two brushed coats of Aeroglaze
Z306. The resulting finish can be seen in Figure 3.13. After curing, this coupon was
subjected to a variety of tests to ensure the coating was well adhered to the copper
substrate. First, it was bent in the middle along its long and short axis multiple times
to see if the paint would flake off or crack. Then, it was pumped down in the vacuum
chamber overnight to see if there was any significant offgassing or if the vacuum would
cause the paint to flake off. After both of these tests, no visible damage to the coating
was present and no significant outgassing was detected, where outgassing would be
indicated by higher than expected pressures and pump down times. Therefore, it
49
Figure 3.12: Second primer coupon with minimal fisheyes and even, thin-
ner application
was determined that this methodology would be adequate for the full size shroud and
heater plates.
3.1.5 Shroud Assembly Painting
After the shroud was brazed and cleaned with citric acid to remove flux and oxides,
the shroud exterior was masked off using blue painter’s tape and a black trash bag.
Because of the known issues with the sprayer causing defects and spraying in a more
speckled pattern instead of an even coating, the primer was diluted before spraying
using thinner as much as allowed per the technical data sheet (TDS). Primer was
then applied using the HVLP sprayer on the inside of the shroud. Finally, a lint free
wipe was used to spread and remove excess primer for an even coating with minimal
fisheyes or other defects. The primer is intended to be applied such that when it
dries it has a dull, matte, see-through appearance [18]. Figure 3.14 shows the shroud
interior after the primer dried, with the masking visible on the outside.
50
Figure 3.13: Paint test coupon after 2 coats of Aeroglaze Z306 on top of
primer
51
Figure 3.14: Masked shroud after primer application
52
After the primer fully dried, the topcoat of Aeroglaze Z306 was applied with the
HVLP per the TDS with the maximum dilution allowed, again to aid in preventing a
more speckled pattern during application. To maintain even and consistent coverage,
some areas were touched up with the HVLP sprayer after the initial passes. After
drying, but within 24 hours of the initial coating, touching up painting using undiluted
Aeroglaze Z306 was performed using a paintbrush to fix minor defects and light spots.
The dried shroud interior after painting can be seen in Figure 3.15. Overall the coating
was consistent with some dust or other debris visible upon close inspection in some
areas. This contamination was not expected to cause any issues with the function of
the coating because it appears the coating envelops the debris, so the function of the
coating is maintained because the surface is still fully coated.
Figure 3.15: Shroud interior after Aeroglaze Z306 application
3.2 Heater Plate Machining
The existing heater strips were tightly fit into the existing slots in the platen and
could not easily be removed without risking damage. Because of this and to save on
53
material cost, the heater plates were made from the existing platen. This required
extensive material removal in order to reduce the thermal mass of the heater plates
compared to the original platen. To begin, the platen was cut into three large plates,
approximately 8”x14”, using a large vertical bandsaw. After rough cutting, the plates
were then cut in half along their thickness and the side without the heaters installed
was not used. Figure 3.16 shows one of three heater plates being cut along its thickness
using a horizontal bandsaw.
Figure 3.16: Heater plate in the horizontal bandsaw to cut along the
thickness
After the thickness was reduced using the bandsaw, the plates were then milled to
their final dimensions using a shell mill to face the remaining thickness and provide
54
a smooth and flat surface for painting. Figure 3.17 shows the sectioned heater plate
with the original cooling pipe cuts visible.
Figure 3.17: Heater plate section before final machining
Due to the long and thin nature of the heater plates, machining required careful
fixturing to minimize chattering during machining, which would leave an undesirable
surface finish. Additionally, because of the volume of material needing to be removed,
many passes were needed to face the plates down on both sides. Figures 3.18 and 3.19
show one of three heater plates after final machining. The dimensions were mesured
to ensure they met the required 0.14” throughout.
After the heater plates were machined, additional 10-32 threaded holes were added
to the corners of the plates to aid in mounting them in the chamber. Additionally,
two strips of aluminum were bolted to the back face of the heater plates to help retain
the heater strips in the channel on the back of the heater plates.
55
Figure 3.18: Heater plate back face after final machining
Figure 3.19: Heater plate front face after final machining
56
Finally, the heater plates were cleaned following the same cleaning process as the
shroud, except citric acid cleaning was not needed since the heater plates did not have
any brazing flux. The channel where the heater strip is installed, visible in Figure
3.18, was masked off using painters tape. Then, primer was applied using lint free
wipes instead of the paint sprayer, because of the smaller area needed and to prevent
possible contamination issues. After the primer dried, the Aeroglaze Z306 topcoat
was applied using a paintbrush and allowed to dry before installing the heater strips.
The heater plates can be seen in Figures 3.20 and 3.21.
Figure 3.20: Heater plates after primer application with additional mount-
ing holes visible
3.3 Modular Platen Assembly
The platen is based on 304 stainless steel oven racks found in conventional home
cooking applications. This was selected as the platen material because it has a large
open area for radiative heat transfer from the upper and lower halves of the chamber,
emissivity under 0.10, favorable out-gassing properties, and low cost compared to
custom sized solutions. Two 10”x15” racks were brazed together along the 15” edge
to form an approximately 20”x15” rack utilizing leftover brazing material from shroud
57
Figure 3.21: Heater plates integrated into the shroud after painting
brazing. Stainless steel tube sections were then brazed to the four corners of the rack
so that it could be easily slid in and out of the chamber along the middle rails. The
final assembly was cleaned using a wire brush, Scotchbrite, and IPA wiping to polish
surfaces oxidized from brazing and to remove brazing flux residues. Figure 3.22 shows
the platen integrated with the shroud. This platen can be easily removed from the
chamber and larger test articles could be mounted using the existing mounting rails
located at 90◦ intervals on the shroud.
Figure 3.22: Modular platen assembly installed on mounting rails inside

shroud
58
3.4 Outgassing Considerations
Every part going into a vacuum chamber must be capable of withstanding the vac-
uum environment without degrading and must not outgas significantly. Outgassing
can cause test pressures higher than required, and condensation of outgassed mate-
rials on sensitive surfaces can cause a variety of issues, such as obscuring sensors,
making non-conductive surfaces conductive, and contaminating thermal control sur-
faces like radiators, which could affect performance. For this project, efforts were
made to ensure the new parts and materials going into the manufacturing process
were adequately cleaned and consisted of vacuum compatible materials appropriate
for achieving a vacuum on the order of 1.0e−6 Torr, with a desired vacuum level of
7.5e−6 Torr as required by ISO 19683 [9].
In general, vacuum compatible materials have an affinity to not adsorb volatile
substances like water vapor, oils, etc. Additionally, every material has a vapor pres-
sure, which can be thought as the pressure of a substance near a liquid or solid
surface of a material in a closed system. In the case of a vacuum chamber where low
pressures are desired, materials with higher vapor pressures will constantly evaporate
into the chamber, increasing the pressure. This pressure can only be reduced by more
pumping, a reduction in temperature, or coating or removing the material entirely.
NASA maintains an outgassing database with information on thousands of dif-
ferent samples [15]. This database contains information on various samples of the
coating used for the shroud and heater plates, Aeroglaze Z306, and the data sheet for
the coating cites a TML of 1.00% [19].
Fermi National Accelerator Laboratory published a technical memo outlining ma-
terials for ultra-high vacuum, which was utilized to help with material selection.
Outgassing of materials is highly non-intuitive, as many metals outgas at significant
59
rates such that they should never be used in a vacuum chamber, like zinc, cadmium,
and brass [13]. Even steels should be degassed since they can contain trapped hydro-
gen that is slowly released into a vacuum system [13]. The Fermi memo is specific
to ultra-high vacuum, and some materials that are mentioned as ”try to eliminate”
like Kapton, Teflon, and zinc [13] alloys are used sparingly in this project by neces-
sity. Additionally, existing wire insulation and plastic connectors are not screened for
compatibility since many consist of unknown materials and have already been shown
to allow the chamber to pump down to a reasonable level.
3.5 Integration
Integrating the new hardware with the chamber required modifications in order to
fit the new shroud in the chamber. The chamber inner diameter is 18”, and exist-
ing mounting rails and hangars protruded into this space, so to fit the 16.5” inner
diameter shroud, these mounting rails and hangars had to be removed. They were
tack welded to the chamber interior, and therefore could be removed by grinding the
weld until they could be freed from the chamber walls. Before grinding, all ports and
orifices were carefully masked off using tape and plastic sheeting so that no debris
would enter where it could not be easily cleaned. This is especially important for the
turbomolecular pump inlet at the back of the chamber, where cleaning would require
extensive tear-down of the pump. After masking, additional material was placed in
the bottom of the chamber to catch most of the grinding debris. After grinding, any
sharp edges or indents in the chamber walls were ground smooth with sandpaper and
Scotchbrite to reduce stress concentrations and prevent injury or damage when slid-
ing the shroud in and out. The chamber was then vacuumed and thoroughly cleaned
with IPA to remove grinding residue and other contaminants.
60
Before test fitting the shroud into the chamber, masking from painting was re-
moved and the mounting rails were installed. Figure 3.23 shows the shroud with
the mounting rails installed. After the mounting rails were installed, epoxy fiber-
glass blocks were mounted to the shroud to support it in the chamber and keep it
thermally insulated from the chamber via direct contact. Two blocks were mounted
underneath the shroud, and one block was mounted on each side, directly adjacent
to the mounting rails and evenly spaced to support the weight of the shroud and
test articles inside. Stainless steel lock wire was used to keep the insulating blocks
attached to the shroud without significant clamping force such that contact resistance
can be maximized to reduce thermal bridging. Each block is approximately 1” x 0.5”
x 0.75” and has a hole drilled through it for the lock wire to pass through. Lock wire
was also used to keep the inlet copper tube that runs from the front of the chamber to
the back of the shroud from contacting the chamber inner walls, which would cause
thermal bridging otherwise.
After the chamber and shroud was prepared for the new hardware, multiple test
fits were made to ensure that the tubing for the coolant was correctly cut and installed.
The tubing terminated with the slot cross section and had to be made circular using
a punch and clamping tube rounding tool so that copper 90 degree fittings could be
brazed on. After the 90 degree fittings, the tubing is brazed to Swagelok stainless
steel caps for 3/8” tubing with a hole drilled in the top of them to allow the coolant
to pass through. Figure 3.24 shows the tubing with the slot cross section cut to the
correct length before bending and brazing.
Because the shroud is already painted and the brazing of the 90 degree fitting
and Swagelok cap is adjacent to the painted surface and connected thermally by
copper tubing, steps were taken to ensure the coating was not overheated during
brazing. Aeroglaze Z306 has a maximum operating temperature of 250◦ F, a wetted
paper towel stack was placed directly in contact with the shroud near the painted
61
Figure 3.23: Shroud after mounting rails were installed
62
Figure 3.24: Trimmed shroud and chamber tubing after test fit
areas next to the tubing being brazed. This keeps the adjacent copper around 212◦ F
through evaporative cooling. Figure 3.25 shows the brazing setup just before brazing
began.
After brazing was complete for the final connections, the flux was removed using
a wire brush, water, and IPA wiping. The shroud was given final IPA wiping before
sliding into the chamber and tightening the Swagelok connectors. These connectors
are designed to be opened and closed repeatedly while maintaining a gas tight seal,
so the shroud can be easily removed if needed by loosening the connectors and sliding
the entire shroud out of the chamber. Figure 3.26 shows the brazed connections,
which must be gas tight to ensure air or nitrogen in the coolant lines does not leak
into the chamber during operation.
After brazing was complete, the heater plates were wired into the existing wiring
and the platen with heater plates on top were slid into the middle mounting rails.
The thermocouple connector plate was mounted to the left middle mounting rail using
stainless steel hardware. Figure 3.27 shows the fully integrated system in the chamber
with thermocouples connected for initial testing.
63
Figure 3.25: Shroud connector brazing setup with wet cooling towels
Figure 3.26: Shroud coolant tubing connections after brazing and fastening
64
Figure 3.27: Shroud and heater plates integrated into the chamber
3.6 Cost and Schedule
The total budget for the modifications, not including liquid nitrogen refills, the up-
graded liquid nitrogen tank and scale, and consumables like acetylene, nitrogen, oxy-
gen, etc. was approximately $700, and was exceeded slightly as more brazing strip
and miscellaneous fittings and screws were needed. For a similar project, the total
modification cost could easily be under $1000 total for a well equipped shop with an
HVLP paint sprayer, mill, bandsaws, ring roller, welding torch, and other hand tools.
The manufacturing portion for the project started with materials purchasing in
early January 2021, and was completed by early May 2021. During this time, approx-
imately 160 person-hours was spent on manufacturing, with the most time consuming
parts being the shroud tubing brazing. This is likely to vary significantly depending
65
on the prior knowledge of the manufacturer as well as their experience with specific
relevant tasks like HVLP spray painting, brazing, and machining.
66
Chapter 4
TESTING AND ANALYSIS METHODOLOGY
This chapter outlines the methodology for the thermal testing and analysis for this
thesis. This includes chamber bakeout, controller configuration, data collection, ther-
mocouple placement, and test configuration for each test as well as pre-test configu-
ration changes to tune the control system. For the analysis methodology, the thermal
model is discussed including validation requirements, model components, material
and surface properties, and assumptions made along with their associated rationale.
A mesh refinement study is presented to demonstrate the mesh size does not signifi-
cantly affect the numerical simulation results.
Figure 4.1 shows the general outline of the methodology for this thesis. This
chapter defines the configuration for the detailed thermal model and thermal vacuum
tests as well as the validation process. Chapter 5 presents the results of these tests
and numerical model correlations.
Figure 4.1: Methodology overview flowchart
67
4.1 Thermal Vacuum Chamber Testing Configuration
4.1.1 Chamber Bakeout
As discussed in the previous section on outgassing considerations, outgassing rates
increase with temperature. This can be used to ones advantage using a chamber
bakeout. A bakeout will reduce outgassing rates before initial testing and is useful
because the Aeroglaze Z306 coating used requires cure time on the order of days
to weeks depending on temperature, humidity, and airflow to reach the 1.0% TML
specified on the data sheet [19]. Bakeouts were performed for this thesis primarily to
accelerate the drying of the Aeroglaze Z306 coating used on the shroud and heater
plates. The total bakeout time above 80◦ C was on the order of days but was not
specifically recorded. Chamber specific bakeout procedures can be found in appendix
A of this thesis.
4.1.2 Controller Configuration
The heating and cooling system is regulated by four Watlow brand controllers, each
with their own specific functions and capabilities. The controller configuration focused
on the Watlow F4DH-KKFK-01 dual channel ramping controller, which was used to
control shroud and platen temperatures. Other controllers were not modified except
to modify the chamber high and low limit alarm temperatures. When the previous
shroud and platen were in use, the temperature was controlled only for the platen,
and the shroud temperatures were correlated with the platen because they shared
nitrogen plumbing lines [3, 11].
Initially, the controller was used as a single channel with the control thermocouple
placed on the center heater plate directly on the heater element. This did not have
the desired effect for chamber control and was modified to allow for independent
68
control. In the initial empty testing section 4.1.6, this is discussed in more detail. For
this thesis and all test results presented the shroud and heater plates were configured
to be independently controlled, with the heaters controlled on CH1 and the shroud
controlled on CH2 of the main Watlow F4 controller.
4.1.3 PID Tuning
The Watlow F4DH-KKFK-01 dual channel ramping controller used to control heater
plate and shroud temperatures is capable of automatically tuning the PID settings
such that the system responds appropriately to a given input temperature and ramp
rate or ramp time. It performs this tuning by ramping to a percentage of the set point
(95% for this testing, based off how it was set previously), and uses on-off control to
move the temperature above and below the tuning set point [21]. After crossing the
set point four times, the controller uses the thermal response of the system to update
the PID values automatically [21]. Figure 4.2 shows the auto-tune profile.
Auto-tuning was used for the heater plates (on CH1) and shroud (on CH2) before
empty test 1. It is not required to re-tune the system between tests unless there
is a significant change in the configuration that would affect the response of the
heaters or shroud during heat-up or cool down. If the system is not responding
appropriately, i.e, it is oscillating around a set point instead of maintaining the value
within +/-1◦ C on the controller input thermocouple, re-calibration can be performed
using the auto-tuning feature or manual tuning. It is recommended to set the auto
tune set point below the maximum/minimum tolerable temperature, since the tuning
process tends to overshoot the set point. It may overshoot if the previous PID values
were tuned for a less responsive system, like the previous platen with larger thermal
mass and greater separation between the strip heaters and the control thermocouple.
Additionally, if the controller behavior is different in different temperature ranges,
69
Figure 4.2: Auto-tuning process of the Watlow F4DH controller [21]
i.e, does not undershoot at 50◦ C but does at 90◦ C, then a separate set of PIDs can
be stored for auto-tuning in different temperature ranges. The controller stores 12
PID sets for both channels, and any PID set can be selected to be used during any
specified ramp or hold [21]. For this thesis testing, one set of PIDs was used for each
channel, PID set 1 for CH1 and PID set 6 for CH2, which is the default or first option
for each channel when setting a profile.
4.1.4 Data Collection
The data collection setup is shown in the annotated Figure 4.3. A tablet computer
with webcam and time lapse software was used to monitor the status of the chamber
and testing remotely as well as record parameters like liquid nitrogen scale weight,
chamber pressure, and the control thermocouple temperatures as well as alarm status
in the case the chamber exceeded the high or low limit alarms.
70
Figure 4.3: Labeled data collection experimental setup
71
4.1.5 Thermocouple Placement
Thermocouples were applied using procedure specified in Appendix B of Jensma’s
thesis [11]. Photos of installed thermocouples can be found in the photos of each test
setup in the sections below along with a labeled diagram of their placement. T-type
thermocouples were used for all data recording and control thermocouples, and have
an error of +/- 1◦ C. Error bars are not included in any plots to improve readability.
4.1.6 Initial Testing
Before empty test 1 started, an initial empty test was performed to learn how the
updated PID values affect the shroud response with the control thermocouples for
the heater and the shroud both on the heating element of the center heater plate. In
the initial testing, the heating and cooling system was controlled by a thermocouple
located on the heater element outer surface of the middle heater plate. This location
means that the temperature of the heater element is precisely controlled, but it does
not correlate to the heater plate outer surface temperatures exactly, and during cool-
ing cycles, the shroud will overshoot to much lower temperature in order to cool the
heater element quickly. Because of these control issues, initial test data is not useful
for analysis and is not presented.
4.1.6.1 Control Thermocouple Placement
Moving the control thermocouple for the heater plate to the top surface of the heater
plate, which is the side exposed to the test article and shroud, allows for the surface
to be maintained at a specific temperature by the heater on the other side of the
heater plate. The second channel of the Watlow F4 controller is then used on the
shroud, between thermocouple 8 and 10 in the empty test thermocouple placement
72
diagram. The PIDs were tuned as discussed in the previous section on PID tuning.
The locations of the control thermocouples and the PID parameters were not changed
between any tests after the initial modification discussed.
4.1.6.2 Coolant Leak Repair
During the initial testing, a leak in the nitrogen inlet line as the line enters the
chamber enclosure was observed. The leak was due to thread tape being applied on
a flared connector fitting, which should not have thread sealant tape since the flared
mating surfaces is intended to provide the leak tight seal instead of the threads. After
repairing the leak by removing the tape and re-fastening the joint, no significant
leakage was visible.
4.1.6.3 Door Shroud Operation
The door shroud was operated during the initial testing and it was determined that
the door shroud consumes significantly more nitrogen than the shroud and provides
a marginal benefit during cold soaks. Additionally, because the door shroud is con-
trolled separately, it cannot be easily turned on and off during the main controller cold
and hot soaks. Similar conclusions were made by Jensma during their experiments,
and they demonstrated the shroud had little effect on test article temperatures [11].
This makes sense because of the limited view factor for test articles with less surface
area normal to the shroud cylinder axis. Despite this, the door shroud functions in
that it is able to cool when nitrogen is flowed through it, and can be used if required,
but was not used for any tests beyond initial testing discussed.
73
4.1.7 Empty Test 1
The purpose of the empty TVAC test is to ensure the chamber is functioning as
expected with regard to following a specified thermal profile and to explore the ca-
pabilities of the modified system, as well as diagnose and adjust the system and
thermal model based on the results. The test includes the cooling shroud, three mod-
ular heater plates, and the platen, as well as 12 logging thermocouples to record the
temperatures across the systems.
4.1.7.1 Thermocouple Placement
Figure 4.4 shows the locations of the thermocouples for both empty test 1 and empty
test 2. Note that shroud thermocouples have a blue outline and heater plate ther-
mocouples have a red outline. All shroud thermocouples are located on the inside
black faces of the shroud, and labels 6, 8, 9 and 11 are projected locations due to
restrictions of the view.
The thermocouples were arranged such that the temperature distribution across
the platen, heaters, and shroud is observable in key locations. The shroud thermo-
couples focus on the middle and front of the shroud since those locations had the
greatest predicted temperature variation based on initial sizing models in Chapter
2. In addition, areas towards the front and back edges of the shroud and the heater
plate middle and edges were monitored to evaluate the maximum and minimum lo-
cations to get an accurate picture of the temperature distribution across the shroud
and heater plate surfaces.
74
Figure 4.4: Empty test 1 & 2 thermocouple placement diagram
4.1.7.2 Experimental Configuration
Figure 4.5 shows the experimental setup for empty tests 1 and 2. It should be noted
that for these tests and test article tests, aluminum foil reflectors were installed in
the back of the chamber, over the door shroud, and over the front face of the shroud
as shown in Figure 4.5 and Figure 4.6. As mentioned in Chapter 2, these allow for
more even temperature distributions in the chamber and preserve cooling and heating
power by reflecting the energy back into the shroud, heater plates, and test article.
4.1.7.3 Thermal Vacuum Test Profile
The thermal vacuum test profile for empty test 1 is intended to allow for two different
temperature ranges to be tested with one cycle per set of hot and low temperatures.
Soak times of two hours were set to allow the chamber to reach steady state. The
75
Figure 4.5: Empty test 1 & 2 configuration experimental setup
Figure 4.6: Empty test 1 & 2 configuration experimental setup with front
foil reflector
76
profile in detail is shown in Table 4.1, including ramp rates and elapsed time. It
should be noted that there is no soak after the first 55◦ C ramp due to user program-
ming error when setting the profile. Because the test demonstrated control meeting
the profile programmed and cycles between high and low temperature ranges with
predicable response, the test results were used and the test was not repeated despite
the programming error.
Table 4.1: Empty test 1 thermal vacuum test profile
4.1.8 Empty Test 2
Empty test 2 shared the same thermocouple placement and experimental configura-
tion as empty test 1 except the thermal vacuum test profile, shown in Table 4.2, was
modified to include more cycles using a low and high temperature range exceeding
the requirement of ISO 19683 for thermal vacuum testing, , which were discussed pre-
viously in Chapter 1 Section 5, by 5◦ C. Soak times were reduced to 2 hours because
empty test 1 demonstrated steady state for the chamber was achieved in less than 2
hours.
4.1.9 CubeSat Mass Model Test
The purpose of the CubeSat mass model test was intended to test the response of a
large test article in the chamber over multiple thermal cycles.
77
Table 4.2: Empty test 2 thermal vacuum test profile
Existing locations for shroud and heater plates were used for locations 1, 4, 6 and
8. The remaining thermocouples were used to measure the temperatures of the test
article. Figure 4.7 shows the thermocouple locations across the test article surface,
with a focus on the front of the test article since that is adjacent to the front of the
chamber, which had more temperature variations than the back of the chamber in
general.
The test article used is a 3U CubeSat mass model. It is made of 6061 Aluminum
with stainless steel fasteners. The surface is anodized and each 1U segment is colored
differently. The test article is shown in Figure 4.8.
The test article was placed in the center of the platen, isolated by Teflon spacers
as shown in Figure 4.9.
78
Figure 4.7: CubeSat mass model thermocouple placement diagram
Figure 4.8: CubeSat mass model test article
79
Figure 4.9: CubeSat mass model test article setup in the chamber
Similar to empty test 2, multiple cycles using a low and high temperature range
exceeding the requirement of ISO 19683 for thermal vacuum testing by 5◦ C were
used. In addition, a hot soak at the beginning allowed for steady state temperatures
for thermal model comparison, and the first cold cycle with a -45◦ C soak was used
to test the response at lower temperatures. Table 4.3 shows the profile in detail.
Note the ramp rate was increased from +/- 1◦ C per minute from the empty tests to
+/- 2◦ C per minute to reduce ramp times and overall test time, with the ISO 19683
requirement of +/- 5◦ C per minute or less. Three hour soaks were used to allow for
more time for the CubeSat mass model to reach steady state, however steady state
was not achieved except for the initial 80◦ C soak.
80
Table 4.3: CubeSat mass model thermal vacuum test profile
4.1.9.4 Aluminum Plates Test
The purpose of the plates test was to further test the response of the chamber at
greater temperature extremes than the CubeSat test to determine the response of
a coated and bare aluminum test article. The test articles used are 145 mm x 102
mm x 6.5 mm 6061 aluminum plates and are shown in Figure 4.10. One plate was
abraded with Scotchbrite to give it a consistent brushed appearance, and the other
plate was primed and painted with Aeroglaze Z306 following similar procedures to
those discussed for heater plate manufacturing in Chapter 3.
81
Figure 4.10: Plates used for testing, one bare aluminum, one painted black
Existing locations for shroud and heater plates were used for locations 1, 4, 6 and
8. The remaining thermocouples were used to measure the temperature of the test
articles. Figure 4.11 shows the thermocouple locations across the test article surface.
The placement of thermocouples for this test is intended to measure in the middle
and edges since the distribution is not expected to be significantly different on any
one side or face. Measuring in the middle allows for more of an average temperature,
and temperatures at the edges typically represent more extremes, based on similar
results with test articles in Chapter 2.
The setup with the plates in the chamber is shown in Figure 4.12. Note that the
plates are placed in the back of the chamber since this part of the chamber has more
consistent temperatures for the surrounding shroud during cold soaks.
82
Figure 4.11: Plates thermocouple placement diagram
Figure 4.12: Plate test articles setup in the chamber
83
The profile includes increasingly colder minimum soak temperatures to test the limits
of the cooling system, with the maximum soak temperatures that the chamber is
capable of at 95◦ C. This represents a final cycle temperature difference of 175◦ C over
a period of 87.5 minutes, thermally stressing the test article and chamber to the
maximum.
Table 4.4: Plates thermal vacuum test profile
4.2 Minimum Temperature Testing
After the plates test was completed, the controller was set to -145◦ C to determine
the lowest temperatures the chamber was capable of achieving. The same setup and
configuration as the plates test was used. The intent of this test was to use residual
nitrogen after testing and allow the chamber to soak at the lowest temperature pos-
sible. The shroud temperature is limited by the coating lower temperature operating
84
limit, which is -150◦ C. There is no profile or additional setup diagrams since the tem-
perature was held constant with an uncontrolled ramp at -145◦ C, and temperatures
were manually recorded.
4.3 Thermal Model Setup
4.3.1 Assumptions and Simplifications
The solid model for this analysis was built up during the modification design and
manufacturing process, which means that the dimensions and materials are generally
well known. However, simplifications and assumptions are still made in the geom-
etry and boundary conditions where necessary. A list of simplifications and their
associated rationale for the thermal model is as follows:
• The chamber exterior is modeled as a cylinder with flat caps on either end with
constant thickness throughout and does not include ports or pass-throughs.
Modeling these accurately would require knowledge of the materials and inter-
faces of the caps, connectors, o-rings, etc. and is not likely to have a impact
on the overall results because the pass throughs are similar in geometry to the
overall cylinder and represent a fraction of the total area of the pressure vessel.
• The shroud supports between the shroud exterior face and chamber interior
face are not included in the thermal model due to their small size and low
thermal conductivity. The total contact area is less than 4 in2 and the material
is an unknown fiberglass reinforced polymer. Similar materials have thermal
conductivity hundreds of times lower than copper, therefore minimizing thermal
bridging.
• The shroud tubing is isolated from the chamber and terminates at the front
of the shroud in the model, where in reality it passes through the chamber
85
walls to the outside. There is likely thermal bridging where the tubing contacts
the chamber to seal the chamber, but the exact geometry is unknown and
not easily accessible, and the tubing is stainless steel after the front of the
shroud, which greatly limits conduction compared to copper. This should have
a minimal impact on the results because the tubing is thermally isolated as
much as possible with foam insulation outside the chamber and seals and low
conductivity materials inside the chamber.
• The shroud wrinkles are not modeled, and the shroud is assumed to be a cylin-
der. This is because modeling the wrinkles accurately would be highly complex
and will not significantly affect the overall results.
• The Aeroglaze coating thickness and material properties are not included in the
model - the surfaces are assigned an emissivity based on the coating emissivity.
This is because the thermal conductivity and thickness of the paint coating is
unknown, but the thickness is on the order of 0.001”, such that the impact on
surface temperatures is expected to be negligible.
• Wires, brackets, and connectors for thermocouples and heater power are not
included in the model. These can act as thermal bridges and can block ra-
diative heat transfer, but are expected to have minimal impact on the overall
model because most wires and connectors are in spaces where they do not sig-
nificantly block radiative heat transfer between the shroud or heater plates and
test articles.
• All contact resistances are assumed to be bonded, i.e, zero thermal contact
resistance. This is because it is difficult to correctly characterize the contact
resistance because of unknown clamping forces and surface qualities. Addition-
ally,the amount of contact between components that are different temperatures
is intentionally limited by design, like contact between the shroud and chamber
86
walls. Assuming bonded contact for components inside the shroud is reasonable
since they are expected to be at similar temperatures during steady state.
• The outer chamber and insulation is assumed to be at a constant 20◦ C bound-
ary condition to simulate ambient conditions. This is not exact but is intended
to approximate the average temperature as it changes due to day/night cycles
since the chamber is in a non-climate controlled lab. It is possible this should
be changed depending on the season to best approximate average ambient con-
ditions. Testing took place in May and June 2021 for this thesis.
• For the CubeSat mass model, simplifications to the geometry were made in-
cluding assuming no bolts or small holes, but major features are maintained
including the outer shell and interior rails. This is acceptable for this type since
emissivity and conductivity largely determines steady state temperatures.
• For simulating hot cases, heater top surface temperature boundary condition is
set based on hot soak set temperature. This is acceptable because the temper-
ature distribution on the heater plate surfaces was measured to be within 1-2
degrees at steady state.
• For simulating cold cases, either a shroud tubing temperature boundary condi-
tion is set based on cold soak set temperature, or flow of nitrogen is simulated
through the shroud tubing using SolidWorks Flow Simulation.
• Material properties and surface properties for some components are assumed due
to lack of verifiable documentation. This is acceptable because the materials
are classifiable by eye and exact alloy details do not greatly affect thermal
conductivity in this case.
87
If the results vary from the thermal model predictions, it may indicate one or more of
these assumptions are not valid and the model will be tuned or detail will be added
until the required predictive capability is met.
4.3.2 Material and Surface Properties
Table 4.5 outlines the material properties used for components in the chamber. Note
that constant properties are not assumed and for materials like aluminum and copper
the conductivity changes significantly with temperature. This is important since the
chamber experiences large temperature swings during cycling.
Table 4.5: Chamber material properties for thermal model
Table 4.6 outlines the surface properties used for components in the chamber along
with the source of that value. Note that the chamber interior walls and shroud outer
surface were user defined based on model tuning that took place after empty test 1.
The thermal model results shown all include the modified emissivities and remain
constant for all simulations shown.
Before the final results for empty test 1, the thermal model was manually tuned
to best match the experimental results. The tuning involved iterating a hot case soak
88
for empty test 1, where the heater plate top surface is set to a constant temperature
boundary condition. The simulation is then run while changing the two user defined
emissivities in Table 4.6 until the modeled temperatures most closely correlated to the
experimental results for empty test 1. This tuning process was manually performed by
changing the user defined emissivities by 0.01 for each run of the model and was only
performed after empty test 1 to tune the thermal model to match the experimental
results, because the exact surface properties for the chamber inner walls and shroud
outer surface is not known as precisely compared to other surfaces. These surface
properties in particular are not known precisely because the inner chamber is not
perfectly rough or polished and the shroud outer surface is not perfectly polished,
compared to a surface like the Aeroglaze coated surfaces where the emissivity is well
characterized by the supplier.
Table 4.6: Chamber surface properties for thermal model
89
4.3.3 Mesh & Mesh Refinement Study
A mesh refinement study was performed to verify heater and shroud average, mini-
mum, and maximum temperatures did not change by more than 1◦ C with mesh size.
The empty model was used for mesh comparisons, and the smallest mesh size was
used for the analysis correlations because all measured temperatures did not change
by more than 1◦ C with mesh size. The larger meshes were approximately three times
larger by total cell count compared to their smaller counterparts. Because there are
experimental results for additional comparison, mesh refinement consisted of only 2
levels for a high level comparison. Simulations with test articles have larger meshes
to include test articles, mesh size is similar in appearance to heater plates depending
on geometry, and always has more total cells because of the addition of the test ar-
ticle. Additional mesh refinement was considered unnecessary for all configurations.
Table 4.7 shows the results for the flow simulation and conduction + radiation only
simulations.
Table 4.7: Mesh refinement study results
90
SolidWorks’ meshing algorithm allows the user to specify a level of detail based
on locations in the model and detail in that region, as well as levels of detail for the
entire model, based on a minimum mesh size and other parameters or based on a
sliding scale. Based on these details, the software calculates the mesh and shows key
parameters like total number of cells.
Figure 4.13 and 4.14 show the meshes used for the empty test comparisons. Note
that the mesh is focused around smaller features, especially the shroud tubing for flow
simulation, and is less refined in less critical areas like the outer chamber. Closeups
outlined in red show the region near the shroud tubing and heater plates in more
detail.
Figure 4.13: Small mesh used for numerical simulations with conduction
and radiation only
4.3.4 Thermal Model Validation
In order to demonstrate that the computational thermal model accurately predicts
temperatures in the chamber, experiments using the modified chamber were per-
formed and compared to the model. The experimental data is considered to be the
91
Figure 4.14: Small mesh used for numerical simulations with conduction,
convection (flow simulation), and radiation
true solution, and the goal is for the computational model to be able to predict
the response of the chamber and test articles accurately before tests are performed.
Whenever a model is referred to, it is referencing the computational models used to
predict the temperature distribution using SolidWorks and SolidWorks Flow Simula-
tion tools as configured and described in this chapter.
For this thesis, two empty tests are performed first with the chamber operating a
specified profile. During these tests, the temperatures at various points of interest are
logged and used to fine tune the thermal model. Fine tuning involves ensuring that
the simplifications and assumptions made to the model are reasonable and result in
accurate predictions of the response of the chamber under the test conditions. After
the model is tuned such that the model accurately predicts the temperatures of the
chamber empty, tests with test articles are conducted and temperatures of the test
article and points of interest in the chamber are recorded. These results are compared
to the numerical model with the test article included.
92
For numerical model comparisons to test results, temperatures are compared at
the same locations as tested on the corresponding computer model, and the absolute
difference in temperature is calculated for each location and reported to the nearest
1◦ C. The maximum absolute difference for each test is reported for all locations or
for a particular component like the shroud or test article. These maximum absolute
differences are used to check against thermal uncertainty margin requirement of 15◦ C,
which is based on the European Cooperation for Space Standardization (ECSS) mar-
gin and is the same margin used in previous thermal model comparisons in Jensma’s
thesis for model vs experimental comparisons [11]. This means that the numerical
model must predict the temperatures at measured locations to within 15◦ C of the
experimental value in all cases to be considered validated.
93
Chapter 5
RESULTS AND ANALYSIS
This chapter discusses the outcomes of leak testing and repair as well as the thermal
vacuum testing results compared to the numerical model results at steady state soak
temperatures where appropriate. Different methods of simulating the cold soaks are
presented, one using only conduction and radiation, and the other simulating nitrogen
coolant flow. The thermal cycle consistency is compared for the CubeSat mass model
testing.
5.1 Leak Testing and Repair
Before modifications were made to the existing heating and cooling system, leak
testing was performed using a vacuum technologies Model 979 Series Helium Mass
Spectrometer Leak detector. This device connects to the chamber in place of the
roughing pump and uses its own pump as well as an inline mass spectrometer to
perform leak detection. Helium from an external tank is manually sprayed on the
outside of the chamber in areas of interest, and if a leak is present, the helium atoms
will enter the chamber and eventually pass through the mass spectrometer in the leak
detector. The leak detector displays the calculated leak rate and creates an audible
tone proportional to the leak rate.
Using the leak detector, one relatively large leak source was identified, which was
the threaded connection where the vent valve connects to the chamber in the back of
the chamber wall on the top left looking from the front of the chamber and is outlined
in red in Figure 5.1.
94
Figure 5.1: Vent valve leak source outlined in red
After it was identified, the valve was removed, cleaned and reinstalled tightly with
teflon tape on the threads. An additional relatively large leak source was identified
as the existing nitrogen plumbing, specifically in the lines running to the platen and
shroud. This leak was not able to be corrected at the time and because the hardware
would be replaced it was ignored. After these leaks as well as other small leaks
were identified and corrected, except for the leak in the existing nitrogen plumbing,
the chamber was pumped down using the roughing and turbomolecular pump to
determine the new baseline pressure. After approximately 23 hours, a vacuum of
3.5e−4 Torr was achieved, which compared to the previous baseline pressure of 5.0e−4
Torr represents a 35% improvement. After the old heating and cooling system was
removed in preparation for the new system installation, an additional pump down test
was performed to determine the effect of removing the old hardware. During this test,
a pressure of 1.5e−5 Torr was achieved, which represents an 80% improvement over
95
the baseline before any leak testing. It was determined that any further leak testing
would likely offer little improvement, so leak testing was considered complete, despite
not meeting the ISO 19683 standard requirement of 7.5e−6 Torr. This is because
the remaining leaks were likely numerous and small, so chasing them all down would
require additional effort that may not be needed after new hardware is installed and
the system is allowed to bakeout. In addition, the new baseline at 1.5e−5 Torr was
achieved in less time than the previous baseline at a higher pressure.
After hardware installation and multiple bakeouts and empty thermal vacuum
tests, the CubeSat mass model test pressure was measured. The control panel with
the ion gauge visible was recorded using a time lapse webcam. Figure 5.2 shows
the pressure recorded during the testing. After approximately 11 hours with the
turbomolecular pump on, chamber pressures were below 7.5e−6 Torr for at least 2.5
cycles between -15◦ C and 50◦ C. The base pressure after the thermal cycling was
complete and after more than 48 hours baking out with the heater plates set to 95◦ C,
the base pressure achieved was 4.4e−6 Torr. Minimum temperature capability testing,
pressures as low as 1.7e−7 Torr were observed.
5.2 Empty Test 1
5.2.1 Experimental Results
The minimum temperatures at the end of the -20◦ C cold case at 9.5 hours was used
to represent steady state for thermal model comparisons. Average temperatures for
hours 25-26 were used for the 80◦ C hot case thermal model comparisons.
The empty test 1 data seen in Figure 5.3 shows the temperature response of the
chamber. Note that the heater plates take longer to reach equilibrium temperatures
in cold cases compared to hot cases. This makes sense because the cooling power
96
Figure 5.2: Annotated chamber pressure during CubeSat mass model test-
ing
comes from the shroud, and has to transfer to the heater plates through radiation
only because there is limited contact between them. The shroud is more responsive,
even in hot cases, because the thin copper sheeting and tubing have low thermal mass
as noted Chapter 2. The response of the heater plates in hot cases lag behind the
profile set temperature, possibly due to worse thermal contact due to the different
thermocouple sizes for the control vs logging thermocouples. Time to reach steady
state in this test, which is time after the soak starts that it takes for the temperatures
to remain constant to within +/- 3◦ C, was less than an hour for the hot cases and
less than 3 hours for the cold case. It is expected to take less time for the hot case
since the shroud coolant takes time to cool down the tubing throughout the shroud
before the entire shroud reaches steady state, whereas the heaters reach steady state
quickly and warm the surrounding shroud quickly.
97
80
60
40
Temperature (°C)
20
HP Mid
HP Left
HP Back
0 HP Front
S Front Mid
S Top Mid
S Bot Mid
-20 S Top Front
S Bot Back
S Left Mid
S Right Mid
-40 Profile
0 2 4 6 8 10 12 14 16 18 20
Time (hours)
Figure 5.3: Logged temperature data during empty test 1
5.2.2 Numerical Model Results
The numerical model results for the empty 1 hot case with flow simulation is shown
in Figure 5.4. The heater plate top surface temperatures are set to 80◦ C based on
the profile set temperature. Note that the heater plate temperatures are not shown
since they reduce the resolution of the legend.
The numerical model results for the empty 1 cold case with flow simulation is
shown in Figure 5.5. As mentioned in the methodology, the blue outlined labels cor-
respond to shroud temperatures and red outlined temperatures correspond to heater
plate temperatures. The shroud tubing is set to a constant temperature boundary
condition of -20◦ C to simulate the coolant through the tubing without requiring the
flow to be simulated.
98
Figure 5.4: Numerical model results for empty test 1, 80◦ C case
99
Figure 5.5: Numerical model results for empty test 2, -20◦ C case without
flow simulation
100
5.2.2.1 Simulating Shroud Coolant Flow
It is possible but impractical to use flow simulation to simulate nitrogen coolant
through the shroud in order to predict the temperatures in the chamber. It is possible
to perform an iterative approach, by iterating different nitrogen flow rates in the
thermal model and running the model until the average temperature corresponds
to the desired controller profile temperature. This method was used to correlate
empty 2 test cold soak temperatures with the thermal model, however it takes orders
of magnitude greater time because multiple calculations must be run with a much
larger mesh for the flow simulation cases. This process can be sped up by ensuring
you are using the results of the previous calculation as an initial condition for the
new simulation. For the empty 2 correlation at -20◦ C, the following conditions were
assumed:
• A flow rate of nitrogen of 1.000 g/s at an inlet temperature of -150◦ C and a
pressure of 21 psi
• A back pressure at the outlet of 5 inches of water with a temperature of -9.5◦ C
(based on measured outlet temperatures during the soak)
• A chamber outer wall temperature boundary condition of 20◦ C
• The coolant flow is assumed to be fully developed, dry gaseous nitrogen
• Gravity is turned on
The numerical model results for empty test 1 cold case with flow simulation is
shown in Figure 5.6. There is a difference in temperature in the shroud temperatures
from front to back, with the shroud minimum and maximum temperatures being
28◦ C in the worst case. This difference is due to the lower flow rate used compared
to the sizing model and because of the single spiral geometry, and is not unexpected.
101
Despite this difference, the average shroud temperature is -20◦ C with the majority of
the shroud area being between -16◦ C and 26◦ C.
Figure 5.6: Numerical model results for empty test 2, -20◦ C case without
flow simulation
5.2.3 Model vs. Experimental Results Comparison
Table 5.1 shows the comparison between experimental and numerical model temper-
atures for the hot case. Note that the shroud front middle and shroud top front,
thermocouples 5 and 8 respectively, have the greatest absolute difference in tempera-
102
tures with consistently higher temperatures in the model vs the experimental results.
This makes sense because front of the chamber has more uncertainty in the geom-
etry due to the more details like the door shroud, gaps between the shroud, door
shroud, and chamber walls, and effects from the inlet and outlet tubing, foil, and
thermocouple wires which are concentrated at the front of the chamber.
Location Location Exp. Temp. Model Absolute

Index Name [◦ C] Temp. [◦ C] Diff. [◦ C]
1 HP Mid. 79 80 1
2 HP Left 80 80 0
3 HP Back 80 80 0
4 HP Front 77 80 3
5 S Front 64 69 5
Mid.
6 S Top Mid. 65 68 3
7 S Bot. Mid. 71 70 1
8 S Top Front 61 67 6
9 S Bot. Back 73 71 2
10 S Left Mid. 65 69 4
11 S Right 66 68 2
Mid.
Table 5.1: Comparison of numerical model and experimental results for

empty test 1, 80◦ C hot case
Table 5.2 shows the results when flow simulation is not used for the cold case.
The cold case with flow simulation is later compared to this cold case without flow
simulation. Similarly to the hot case, the front of the shroud is least correlated
with the model, but this time the temperatures are consistently lower in the model
compared to the experimental results. This suggests some aspect of the front of the
chamber is not being precisely modeled enough to produce similar differences as other
locations, or experimental setup error. This could be due to a difference in shroud
or platen surface temperatures, or the reflectors performing worse than expected,
among other possible reasons like the geometry differences discussed in the previous
paragraph.
103
1 HP Mid. -15 -19 4
2 HP Left -15 -19 4
3 HP Back -19 -17 2
4 HP Front -9 -16 7
5 S Front -10 -20 10
Mid.
6 S Top Mid. -17 -20 3
7 S Bot. Mid. -15 -20 5
8 S Top Front -11 -20 9
9 S Bot. Back -22 -20 2
10 S Left Mid. -18 -20 3
11 S Right -18 -20 2
Mid.

empty test 1, -20◦ C cold case without flow simulation
Table 5.3 shows the same cold case experimental results compared to the model
with flow simulation results. This simulation has a 50% decrease in maximum absolute
difference compared to the model without flow simulation. This indicates that the
flow simulation improves the predictability for cold cases, at the cost of more effort
computationally. As an example, iterating the flow simulation model and solving it
many times takes on the order of days whereas the model without flow simulation
takes minutes to run on the same computer. The shroud front middle and shroud top
front continue to be sources of greater difference compared to the average, however
this is less of the case when flow simulation is included since two additional locations
have 5◦ C absolute differences. This greater difference is likely due to the same reasons
as mentioned in Section 5.2.3.
104
1 HP Mid. -15 -18 3
2 HP Left -15 -18 3
3 HP Back -19 -22 3
4 HP Front -9 -14 5
5 S Front -10 -15 5
Mid.
6 S Top Mid. -17 -18 1
7 S Bot. Mid. -15 -18 3
8 S Top Front -11 -16 5
9 S Bot. Back -22 -27 5
10 S Left Mid. -18 -18 1
11 S Right -18 -20 2
Mid.

empty test 1, -20◦ C cold case with flow simulation
5.3 Empty Test 2
Figure 5.7 shows the experimental results for empty test 2. Note the consistency
between the thermal cycles, with similar, repeatable response when comparing iden-
tical cycles. Some variation is seen during the hot soaks and is most visible in the
heater plate temperatures. This is likely variation due to the heater PID controller
as it attempts to reach equilibrium without over or undershooting the target value
significantly. For cold soaks, the front of the shroud tends to be warmer than the
middle and back of the shroud.
Empty 2 numerical model results at -20◦ C are the same as the results presented
in empty test 1 Figure 5.5 without flow simulation because they share cold soak
temperatures. The results for the 55◦ C hot case are shown in Figure 5.8. Similarly to
105
60
HP Mid
HP Left
50 HP Back
HP Front
S Front Mid
40
S Top Mid
S Bot Mid
30 S Top Front
S Bot Back
Temperature (°C)
S Left Mid
20 S Right Mid
Profile
10
-10
-20
-30
0 2 4 6 8 10 12 14 16
Time (hours)
Figure 5.7: Logged temperature data during empty test 2
empty test 1 hot cases, the heater plate top surfaces are set with a constant surface
temperature equal to the profile set temperature, with the difference being empty test
1 compared at 80◦ C instead of 55◦ C because it had a longer dwell at 80◦ C in empty
test 1.
The empty 2 experimental results compared for the -20◦ C case uses average of the
lowest temperatures at the end of each cold soak, and the hot case uses an average
temperatures at hours 12.5-12.7 during the last hot soak for the steady state compar-
isons. Comparisons for the hot case are shown in Table 5.4 with a similar maximum
absolute difference to the empty 1 hot case results with similar higher temperatures
than predicted in the model vs the experimental case.
106
Figure 5.8: Numerical model results for empty test 2, 55◦ C case
107
1 HP Mid. 54 55 1
2 HP Left 55 55 0
3 HP Back 54 55 1
4 HP Front 55 55 0
5 S Front 41 48 7
Mid.
6 S Top Mid. 48 48 0
7 S Bot. Mid. 50 49 1
9 S Bot. Back 50 49 1
10 S Left Mid. 47 48 1
11 S Right 48 48 0
Mid.

empty test 2, 55◦ C hot case
Similarly to empty test 1, the results for the -20◦ C cold case are compared with
and without flow simulation in Tables 5.6 and 5.5 respectively. Similar trends to
empty test 1 in locations and magnitudes of absolute temperature difference are seen
comparing the model and experimental results.
5.4 CubeSat Mass Model Test
After the empty tests were completed and the results demonstrated that the chamber
was capable of consistent control to a specified thermal profile for multiple cycles,
test article testing was performed to compare the thermal response of the test article
for various test profiles. The empty test results demonstrated a maximum absolute
difference of 11◦ C between the predicted and measured temperatures at all measured
locations, indicating that the chamber is consistently predicable within the required
15◦ C thermal uncertainty margin.
108
1 HP Mid. -19 -19 1
2 HP Left -18 -19 1
3 HP Back -21 -17 4
4 HP Front -15 -16 1
5 S Front -9 -20 11
Mid.
6 S Top Mid. -22 -20 2
7 S Bot. Mid. -19 -20 1
8 S Top Front -14 -20 6
9 S Bot. Back -26 -20 6
10 S Left Mid. -23 -20 3
11 S Right -23 -20 3
Mid.

empty test 2, -20◦ C cold case without flow simulation
The CubeSat mass model subjected to the previously specified thermal vacuum
test profile in the chamber and the experimental results are compared to the thermal
model results. Additionally, the consistency of identical cycles is compared for the
CubeSat mass model test to compare the response of the chamber and test article
after identical thermal cycling.
The CubeSat mass model test results are shown in Figure 5.10. Note that the test
starts from a high temperature soak which was the end of a pre-test bakeout. It was
included in the results because the initial 80◦ C hot soak is compared to numerical
model results. The remaining soaks and dwells never achieved steady state for the
test article temperatures, as indicated by the slope of the temperatures before each
soak ends, therefore no numerical model correlations are made for the cold case.
109
1 HP Mid. -19 -18 1
2 HP Left -18 -18 0
3 HP Back -21 -22 1
4 HP Front -15 -14 1
5 S Front -9 -15 6
Mid.
6 S Top Mid. -22 -18 4
7 S Bot. Mid. -19 -18 1
8 S Top Front -14 -16 2
9 S Bot. Back -26 -27 1
10 S Left Mid. -23 -18 5
11 S Right -23 -20 3
Mid.

empty test 2, -20◦ C cold case with flow simulation
CubeSat mass model results for the 50◦ C hot case are shown in Figure 5.10. Similarly
to empty tests hot cases, the heater plate top surfaces are set with a constant surface
temperature equal to the profile set temperature. Note that the CubeSat surface
temperature does not vary by more than +/- 1◦ C at steady state in this case.
Comparing the absolute differences for the CubeSat mass model hot case, Table 5.7
demonstrates a maximum absolute difference of 3◦ C for all locations and a maximum
difference of only 2◦ C for the CubeSat at the measured locations. This indicates that
the thermal model is capable of accurately predicting the CubeSat surface temper-
atures in the hot case. It makes sense that the CubeSat surface temperatures have
a lower absolute difference than the shroud because the CubeSat is more isolated
thermally since it can primarily only be affected by radiation from the shroud, heater
110
80
60
40
Temperature (°C)
20
-20
-40 HP Mid 3
HP Front 5
S Top Mid 9
S Top Front 10
-60 S Right Mid 12
2 Profile
-80
5 10 15 20 25
Time (hours)
Figure 5.9: Logged temperature data during CubeSat mass model test
plates, and reflectors, and it sees more of an average environment temperature than
the shroud or heater plates alone.
5.5 Aluminum Plates Test
The experimental results for the aluminium plates test are shown in Figure 5.11. Note
that the thermocouple numbers with a P refer to the painted plate and thermocouple
numbers with a U refer to the unpainted plate. This test went to the limits of the
chamber’s temperature capability. Despite the dwell times being 2 hours, the plates
did not reach steady state, which is especially true for the hot cases with the unpainted
plate as indicated by the greater slope of the unpainted plate temperature response
at the end of each soak.
111
Figure 5.10: Numerical model results for CubeSat mass model test, 80◦ C
case
112
1 HP Mid. 79 80 1
4 HP Front. 77 80 3
6 S Top Mid. 68 68 0
11 S Right Mid 69 69 0
2 See Dia. 69 69 0
3 See Dia. 69 70 1
5 See Dia. 71 69 2
9 See Dia. 70 70 0
10 See Dia. 69 70 1
12 See Dia. 70 70 0

CubeSat mass model, 80◦ C hot case
Steady state was not achieved for the CubeSat mass model during the test because
the dwell period was too short. As a result, the thermal model results could not be
compared for any soaks during the test. To allow for a thermal model comparison,
after the test was completed the chamber was set to dwell at 90◦ C. Results for hours
9-10 of the soak which were used for the comparison and are shown in Figure 5.12
Aluminum plate model results for the 90◦ C hot case are shown in Figure 5.13. Sim-
ilarly to empty tests hot cases, the heater plate top surfaces are set with a constant
surface temperature equal to the profile set temperature. Note that the plate surface
temperatures does not vary by more than +/- 1◦ C for each plate at steady state in
this case.
Comparing the absolute differences for the aluminum plates hot case, Table 5.8
demonstrates a maximum absolute difference of 3◦ C for all locations and a maxi-
113
100
80
60
40
Temperature [°C]
20
0 HP Mid
HP Front
3P
-20 5P
7P
11P
-40 S Top Mid
S Top Front
2U
-60 9U
10U
-80 12U
0 2 4 6 8 10 12 14 16 18
Time [hours]
Figure 5.11: Logged temperature data during aluminum plates test
90
88
86 HP Mid S Top Mid

HP Front S Top Front
3P 2U
Temperature [°C]
5P 9U
84 7P 10U
11P 12U
82
80
78
76
9 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 10
Time [hours]
Figure 5.12: Logged temperature data during hot soak after aluminum
plates test
114
Figure 5.13: Numerical model results for plates test, 90◦ C case
115
mum difference of 3◦ C for the plates at the measured locations. This indicates that
the thermal model is capable of accurately predicting the plate surface temperatures
in the hot case with the maximum difference well within the required 15◦ C thermal
uncertainty margin.
1 HP Mid. 89 90 1
4 HP Front. 89 90 1
6 S Top Mid. 79 76 3
3 Painted 82 79 3
5 Painted 81 79 2
7 Painted 81 79 2
11 Painted 81 79 2
2 Unpainted 80 80 0
9 Unpainted 81 80 1
10 Unpainted 80 80 0
12 Unpainted 81 80 1

the plates test 90◦ C hot case
5.6 Thermal Cycling Consistency
The consistency of identical thermal cycles is compared by calculating the absolute
difference in temperature for the last two thermal cycles of the CubeSat mass model
test. The temperature response for each cycle is overlayed and subtracted from each
other and the absolute difference is plotted vs time and shown in Figure 5.14. The
cycles compared take place approximately from hours 10.0 to 24.5 and hours 11.0 to
17.5. Note that there is a greater absolute difference during cold transients than hot
transients. This makes sense because the heaters are simply controlled with consistent
power per time on, while the cooling power for the nitrogen system depends on more
variables including external tank pressure, inlet line temperatures, variation in fluid
flow, etc. This comparison demonstrates that from hours 19 to 24, for steady state
116
dwells where no cold transients take place, that the chamber systems can maintain
CubeSat mass model surface temperatures to within 1.5◦ C between each cycle at all
times.
Figure 5.14: Annotated plot of absolute difference in CubeSat temperature

response for last 2 cycles of test
5.7 Thermal Model Correlation Summary
Table 5.9 shows the maximum absolute difference between the numerical model results
and experimental results. The worst case for all cases is a difference of 11◦ C. The
difference not including cold soaks without flow simulation drops to 7◦ C and the
difference for test article temperatures only is 3◦ C. Hot cases are more predictable
than cold cases with or without flow simulation included. This makes sense because
the cold cases have to simulate the nitrogen through the shroud tubing in order to
117
precisely determine the temperature distribution or otherwise approximate it with a
constant temperature boundary condition on the tubes whereas the hot cases just
assume a constant surface temperature boundary condition for the top of the heater
plates.
Test Case Max.

Name Abs.
Diff.[◦ C]
Empty 1 80◦ C hot 6
Empty 1 -20◦ C cold w/o flow simulation 10
Empty 1 -20◦ C cold w/ flow simulation 5
Empty 2 55◦ C hot 7
Empty 2 -20◦ C cold w/o flow simulation 11
Empty 2 -20◦ C cold w/ flow simulation 6
CubeSat 80◦ C hot 3
Plates 90◦ C hot 3
Table 5.9: Comparison of maximum absolute difference between numerical

model and experimental results for all tests
118
Chapter 6
CONCLUSION AND FUTURE WORK
6.1 Conclusion
The goals of the modifications and testing were to improve and document the ca-
pabilities of the Blue Thermal Vacuum Chamber, the conclusion section will outline
the effects of the modification as it relates to ISO 19683 and other requirements,
referencing the testing and analysis as evidence for each capability. Then, the accu-
racy and practicality of the thermal model will be discussed. Figure 6.1 shows the
state of the chamber in its previous configuration compared to the chamber with the
modifications from this thesis. This figure shows the increase in available test volume
and highlights the differences between the configurations, including the addition of
coatings, less thermally massive materials, and improved modularity.
Figure 6.1: Chamber modifications (right) shown with previous configu-

ration (left) [11]
119
6.1.1 Temperature Range
Testing for this thesis demonstrated the chamber was capable of heater plate temper-
atures of 95◦ C and corresponding shroud temperatures >75◦ C and aluminum plate
temperatures of >75◦ C at the hottest case. Temperatures as low as -145◦ C were mea-
sured at the shroud control thermocouple location during extended cooling testing,
and temperatures of approximately <-55◦ C for the entire shroud and <-70◦ C for the
aluminum plates during aluminum plate testing. Previous testing by Jensma was per-
formed with controller temperatures set between -15◦ C to 50◦ C. Therefore, the testing
in this thesis demonstrates capability well beyond the -15◦ C to 50◦ C requirement of
ISO 19683.
6.1.2 Chamber Pressure
Based on the leak testing and repair in chapter 4 demonstrated that after significant
chamber baking out and a low-outgassing test article are used, pressures exceeding
the ISO 19683 requirement can be met after a pump down time of approximately
11 hours. Additionally, pressures as high as 10e−4 are allowed by MIL-HDBK-340A,
TR-2004(8583)-1 Rev.A., and NASA LSP-REQ-317.01 Rev.B. [4]. These higher base
pressures were achieved within minutes after the turbomolecular pump is powered
on and was maintained throughout all tests. Compared to the lowest pressure from
previous testing by Jensma of 2.0e−4 this represents an improvement of 2 orders of
magnitude [11]. Therefore, this thesis demonstrates the system is capable of meeting
and exceeding ISO 19683 chamber pressure requirements for thermal vacuum testing
and the pressure is sufficiently low to assume that convection is negligible.
120
6.1.3 Number of Cycles
The maximum number of cycles for any given test was four during the CubeSat mass
model test, which exceeds the ISO 19863 minimum requirement of 2. The chamber
is capable of more cycles than that per tank of liquid nitrogen. For example, all of
the testing for this thesis required 2 refills of the 180 L tank including time running
extended cooling testing to determine minimum temperatures, as well as calibration
testing that consumed significant amounts of nitrogen at the beginning of testing.
Based on the tests performed, it is estimated that more than 10 cycles between -
15◦ C and 55◦ C could be performed with 2-3 hour dwells using a single tank of liquid
nitrogen. Consumption during the test can be estimated by recording the scale with
a time lapse camera or webcam and calculating the average mass decrease of the tank
during ramps and soaks. These values can be used to estimate the number of cycles
that the chamber will be capable of performing for any given test setup and test
profile. For comparison in previous testing by Jensma, only one complete thermal
cycle was able to be completed due to the limitations of the previous hardware and
35 L liquid nitrogen dewar [11].
6.1.4 Tolerance Limits and Temperature Variation
During test article testing, it was demonstrated that the test articles were able to
maintain surface temperature at the measured locations within a 3◦ C band at steady
state. This means that the chamber cooling and heating systems, on average, are able
to maintain the surface temperatures of these test articles with consistent material
and surface properties. Additionally, cycles were compared with each other for the
CubeSat mass model test and demonstrated less than 1.5◦ C difference in CubeSat
surface temperatures from one cycle to another. However, the chamber is not capable
of achieving test article temperatures within a 3◦ C band of the set temperature in the
121
steady state. The disagreement between the controller set temperature and the test
articles could be due to different thermocouple setups for the control vs data logging
thermocouples. As is mentioned in the future work section, the control thermocouples
are smaller and more flexible and therefore able to be in better contact with surfaces
compared to the thicker thermocouples.
6.1.5 Thermal Dwell
The chamber is capable of dwells of 1 hour or greater as required by ISO 19683 as
demonstrated in all tests. It is expected that extended dwells beyond the 0-3 hour
range tested can be performed as desired with multiple cycles at each dwell. For this
testing, shorter dwells were used in general to conserve liquid nitrogen to minimize
refills needed. The previous system as tested by Jensma could only dwell for an hour
at -15◦ C due to a combination of the limitations of available liquid nitrogen, uncoated
surfaces, and high thermal mass [11].
6.1.6 Temperature Ramp Rate
The ramp rate used during all testing for this thesis varied from +/- 1 to 2◦ C, which
meets the ISO 19683 requirement of +/- 5◦ C or slower. It is possible that faster
ramp rates are achievable and are desirable when the test article is capable of such
ramp rates. This is because faster ramp rates save heating and cooling power by
reducing the total test time. The reason they were not tested up to +/- 5◦ C for this
test was to minimize overshoot. Minimizing overshoot is important because it can
damage the test article or chamber if the maximum or minimum temperatures for the
materials are exceeded, and because the high limit alarm may trigger, which shuts
off the relevant heating or cooling system, interrupting the test.
122
In previous testing by Jensma, programmed ramp rates of -5◦ C/min were mea-
sured at approximately -0.2◦ C/min, and a maximum of -0.9◦ C/min when manually
stepped in temperature, whereas in this thesis the programmed rates of -1◦ C/min
and -2◦ C/min were maintained [11]. This demonstrates an increase in control pre-
cision with the new configuration, however in both the ISO 19683 requirement of
+/-5◦ C/min or slower.
6.1.7 Test Article Volume
With a shroud inner diameter of 16.5” not including mounting rails, the available test
volume increased, and the current system is capable of fitting CubeSats 6U or larger
with the existing modular platen without modifications. The system in its previous
configuration had approximately half the usable test volume [11].
6.1.8 Thermal Model Correlation
The thermal model and experimental correlations demonstrate that the thermal
model is validated with a thermal uncertainty margin requirement with less than
the maximum of 15◦ C in all cases, and as low as 3◦ C in the best case, keeping in
mind thermocouple measurement error alone is +/- 1◦ C.
While the previous experimental setup and analysis conducted by Jensma is not
directly comparable due to geometry and setup differences as well as differences in how
the the thermal uncertainty margin is calculated, the values can be roughly compared
for the cold and hot soaks. The platen temperature thermal uncertainty margin from
Jensma during cold soaks and hot soaks was 9.5◦ C and 7.2◦ C respectively, and the
CubeSat mass model temperature thermal uncertainty margin during cold and hot
soaks was 1◦ C and 11◦ C respectively [11]. In this thesis, the worst case thermal
uncertainty margin for the shroud and heater plates during cold soaks and hot soaks
123
was 11◦ C and 7◦ C respectively, and the CubeSat mass model thermal uncertainty
margin during hot soaks was 2◦ C.
6.2 Lessons Learned
There were a variety of lessons learned throughout this thesis, and they are noted
below and organized by chapter, including modification design, manufacturing pro-
cesses, thermal vacuum testing, and results and analysis.
6.2.1 Modification Design
The main lessons learned during the modification design process is to be sure to add
margin to the design, and to try as much as possible to design to the right conditions.
Specifically, during the modification design the liquid nitrogen flow rates used to check
the temperature distributions were estimated based on previous testing observations.
These estimated flow rates on the surface seemed acceptable, but if other conditions
with less or more flow was modeled and evaluated, a different shroud design may have
been more suitable for this application. Specifically, the flow used for checking the
temperature distribution was higher than what ultimately occurred during testing,
which meant that the temperature distribution was greater than expected during cold
soaks.
Another major lesson was to make sure to keep track of geometry and how all the
parts fit together carefully. Initially the shroud sheet was sized with a 16.5” diameter
to fit in the chamber with barely enough clearance for the L-shaped bars welded
inside the chamber that held up the old hardware. However, with the addition of the
tubing during the brazing process and due to workmanship, the final diameter was
not exactly 16.5” and was barely too large to fit in the chamber with the L-shaped
bars. The bars had to be removed, which added to the manufacturing time, and
124
required masking and cleaning the chamber before and after removing the bars to
prevent contamination of the inside of the chamber and the turbo molecular pump.
6.2.2 Manufacturing Processes
The manufacturing process for this thesis ended up being a significant time commit-
ment, with manufacturing starting in mid-January 2021 and finishing by early May.
From the authors personal experience, manufacturing always takes longer than one
expects, but this ended up significantly exceeding that. A major lesson learned from
this is to plan as detailed as possible, each process and step that is required to com-
plete something in order to properly assess how long it will take, and add margin to
that.
Unfamiliar processes like brazing over 40 feet of tubing to the copper sheet added
significant troubleshooting time spent determining the most efficient methods to com-
plete the work. Coupon level testing is critical for unfamiliar process, and generally
should be used for any process as practice for making the final part if small volumes
are needed and mistakes that could scrap the entire part are possible. The coupon
testing performed for this thesis was extremely valuable, especially for troubleshooting
painting issues and practicing brazing techniques.
6.2.3 Thermal Vacuum Testing
The most important parts of testing are to ensure that the conditions are consistent
and the user completely understands all the systems and how they operate together.
One issue when initially setting up the data-logger is that when it powers on it
does not default to T type thermocouple input, and the default produces similar but
slightly incorrect temperature readings. This is easy to miss before a test and one
125
can change it during the test, but it invalidates the data up to that point unless it
can be corrected for.
Understanding the system completely is critical, especially when it comes to con-
trol systems. Significant time was needed reading through the manual for the PID
controller so that the control system could be setup as desired.
6.3 Future Work
Future work includes potential changes that could be made to further improve the
capabilities or functionality of the chamber, as well as potential test types that could
be performed to further demonstrate the capabilities of the chamber.
6.3.1 Powered Test Article
There are various feed-throughs and connectors with the chamber that could be used
to feed power or data to a test article during thermal vacuum testing. For high power
test articles, black wires labeled 3 and 4 in the back right of the chamber are spare
heater power wires and are currently not connected on either end. These could be
used to feed power or data through to a test article easily with an external power
supply or other equipment. A powered test article test would allow for capabilities like
thermal balance tests with active spacecraft systems, where thermal control systems
on the spacecraft are used to maintain operational limits. Alternatively, powered
subsystems could be tested using a similar experimental setup.
6.3.2 Independent Heater and Shroud Control Tests
As configured, the shroud is capable of being independently controlled from the
heaters. This allows the shorud to act as a radiative sink similar to the cold dark-
126
ness of deep space. The heater plates can then be mounted as desired to simulate
heat input from one or multiple directions depending on how they are arranged. The
plates can be moved as needed and not all plates are required to be in the chamber
at once, and are sized to be slightly larger than the largest face of a 3U CubeSat.
This configuration is expected to consume more liquid nitrogen because the heat in-
put from the heaters will require more shroud cooling. One way to mitigate this is
to wrap the bottom face of heater plate, the side not facing the test article, in high
vacuum compatible aluminum foil to reflect heat away from the shroud and back to
the heater face.
6.3.3 Leak Testing
Ongoing work is needed to correct remaining smaller sources of leakage as well as
perform maintenance if required. The door gasket, as mentioned in Jensma’s thesis
[11], remains a likely source of leakage and will require regular maintenance as it
degrades over time and with usage. Feed-throughs should be checked in more detail
using the helium leak detector. This can be accomplished by using an enclosure like
a plastic bag or other container that focuses the helium injected to specific parts of
the chamber in order to more precisely locate remaining leak sources. This was done
only to a limited extent during the leak testing for this thesis.
6.3.4 Heater Modifications
Currently, the heater plates act as the only source of heat for the chamber, and
because of the limited size and number, are not able to provide a perfectly even
temperature distribution across the entire shroud. Generally, the heater plate tem-
peratures on average were 10-15◦ C hotter than the rest of the shroud. Depending
on the requirements and needs of a specific test, this may be acceptable. However,
127
if it was desired to have an even temperature distribution across the entire shroud
for a test, strip heaters can be installed on the shroud exterior faces. Thin, flexible
polyimide strip heaters are available from suppliers like Watlow for high vacuum ap-
plications that could be bonded to the exterior faces of the shroud to heat the entire
shroud during hot soaks. This system could utilize the existing power feed-through for
the heaters without modification if they are 240V compatible and would not require
permanently removing the heater plates, since they can be unplugged and removed
from the chamber without any tools. This would allow the chamber to be heated and
cooled entirely by the shroud, increasing the available test volume and allowing for
more consistent environmental temperatures for test articles during hot soaks.
6.3.5 Cooling System Improvements
While the upgraded cooling system improved the chamber capabilities, there is al-
ways room for improvement, especially with regard to the temperature distribution
across the shroud during cold soaks. Temperatures across the shroud varied by ap-
proximately 16◦ C in the worst case at steady state during empty test 2. This is not
surprising because the model with flow simulation predicted similar variation, but
could affect the temperature distribution of test articles depending on the test re-
quirements. In order to improve the temperature distribution across the shroud, a
variety of solutions could be tested, either experimentally or through the model, to
improve the distribution. Possible changes include but are not limited to:
• Limiting the maximum flow rate from the liquid nitrogen tank to allow the
chamber valves to spend less time cycling on and off and act more as continuous
flow
• Modify the inlet and outlet conditions by changing the inlet pressure and/or
outlet backpressure
128
• Modifying the outlet geometry by adding orifices to modulate maximum mass
flow rates
• Installing a sensor to measure flow rate
Many of these modifications can be simulated using the model with flow simulation,
and then the changes could be implemented on the physical system.
6.3.6 Data Acquisition System
The existing T type thermocouples and DT4208SD 12 channel thermocouple logger
are adequate for testing, however the logger is currently battery operated and requires
replacement after about few days of continuously logging data. The logger has a 9V
input for powering from a standard 120V outlet, and this could be acquired so that
batteries are not needed and there is no risk of loss of battery power causing loss of
data.
The T-type thermocouples used inside the chamber for data logging have thick,
difficult to bend wires compared to the T-type control thermocouple thermocouples.
The thickness and lack of compliance means that when applying the thermocouples,
it is difficult to ensure good thermal contact between the thermocouple and surface,
and there is a greater risk of thermocouples lifting from the surface and not providing
good contact. It was noted during testing that the control thermocouples tended
to report temperatures closer to the set point compared to thicker thermocouples
directly adjacent to the control thermocouples, and tended to more quickly reach
steady temperatures due to their smaller thermal mass and better contact. It is
recommended that new thermocouples be purchased that are compliant and have
thinner wires.
129
6.3.7 Cooling System Insulation
Before testing for this thesis, the cooling system inlet tubing from the tank to the
chamber was improved using standard polyethylene pipe insulation available at hard-
ware stores. This insulation was wrapped around existing degraded tubing where it
was present, and was added along the hose that runs from the liquid nitrogen tank
to the chamber. Despite these additions, there are more locations where insulation
would help reduce liquid nitrogen consumption and are shown in Figures 6.2 and 6.3.
In Figure 6.2, multiple locations around the liquid nitrogen tank top show icing, in
addition to the valve stem and exposed tubing for the liquid withdraw lines. Insulated
covers could be made for these exposed parts using a soft and flexible type of insu-
lation material rated for cryogenic temperatures which will improve cooling system
efficiency.
Figure 6.2: Insulation around the liquid nitrogen tank inlet and uninsu-
lated areas with ice buildup
Figure 6.3 shows the cryogenic valves uninsulated with significant ice buildup.
These could similarly be covered with a soft and flexible insulation material rated for
cryogenic temperatures.
130
Figure 6.3: Insulation around chamber cryogenic valves and uninsulated
areas with ice buildup
6.3.8 Webcam
It would be useful to be able to view the control panel of the chamber as well as
the liquid nitrogen tank scale and thermocouple logger for the purpose of data log-
ging and monitoring. For example, using a webcam with a time lapse allowed for
chamber pressure to be logged during overnight tests. For this thesis, a personal
tablet computer with a camera was used and directed at the instruments. Installing
a permanent webcam using a computer or similar setup would allow any students to
access it easily for their testing and would not require personal resources.
131
BIBLIOGRAPHY
[1] General Environment Verification Standards (GEVS) for GSFC Flight
Programs and Projects. Technical Report GSFC-STD-7000A, NASA
Goddard Space Flight Center, Greenbelt, Maryland, Apr. 2019.
[2] BasicCopper. 24 X 6’/ 5 Mil (36 ga.) Copper Sheet.
https://basiccopper.com/24x65micofo1.html, 2021.
[3] M. Caudill and B. Diamond. Cal poly tvac chamber: Cryogenics. 2019.
[4] C. D. O. L. Chisabas R. S. S., Loureiro G. and C. D. F. Method for cubesat
thermal-vacuum cycling test specification. In 47th International Conference
on Environmental Systems, Charleston, South Carolina, 2017.
[5] Copper Development Association, Inc. Cryogenic properties of copper.
https://www.copper.org/resources/properties/cryogenic/, 2021.
[6] N. A. Fishwick, K. A. Smith, and J. A. Romera Perez. Lessons learned from
thermal vacuum testing of lisa pathfinder over three system level thermal
tests. In 45th International Conference on Environmental Systems,
Bellevue, Washington, 2015.
[7] D. G. Gilmore. Spacecraft Thermal Control Handbook. American Institute of
Aeronautics and Astronautics, 2002.
[8] J. L. Golden. Results of the examination of ldef polyurethane thermal control
coatings. Technical Report NASA Contractor Report 4617, Boeing Defense
& Space Group, Seattle, WA, July 1994.
132
[9] International Organization for Standardization, Geneva, Switzerland. ISO
19863: Space systems — Design qualification and acceptance tests of small
spacecraft and units, 2017.
[10] S. Jayaram and E. Gonzalez. Design and construction of a low-cost economical
thermal vacuum chamber for spacecraft environmental testing. Journal of
Engineering, Design and Technology, 9:47–62, 2009.
[11] M. R. Jensma. Thermal modeling and testing of the blue thermal vacuum
chamber. Master’s thesis, California Polytechnic State University, San Luis
Obispo, San Luis Obispo, CA, June 2020.
[12] R. D. Karam. Satellite Thermal Control for Systems Engineers. American
Institute of Aeronautics and Astronautics, 1998.
[13] G. Lee. Materials for ultra-high vacuum. Technical Report TM-1615, Fermi
National Accelerator Laboratory, P.O. Box 500, Batavia, Illinois, Aug. 1989.
[14] McMaster-Carr. Cleaned and capped copper tubing.
https://www.mcmaster.com/5174K21/, 2021.
[15] NASA. Outgassing data for selecting spacecraft materials.
outgassing.nasa.gov, 2021.
[16] Paul Fredericks, Aerospace Metals LLC. How to remove anodized aluminum?
https://aerospacemetalsllc.com/how-to-remove-anodized-aluminum,
2021.
[17] V. L. Pisacane. The Space Environment and Its Effects on Space Systems.
American Institute of Aeronautics and Astronautics, 2008.
[18] Socomore. Aeroglaze 9947 Wash Primer Technical Data Sheet. Socomore by
Lord Corporation, Saegertown, PA, 2021.
133
[19] Socomore. Aeroglaze Z306 Polyurethane Coating Technical Data Sheet.
Socomore by Lord Corporation, Saegertown, PA, 2021.
[20] S. Team. SolidWorks Flow Simulation Manual. Dassault Systèmes SolidWorks
Corporation, Waltham, MA, USA, 2020.
[21] Watlow Inc. Watlow Series F4S/D User’s Manual. Watlow Winona, Winona,
Minnesota, 2011.
[22] A. Williams. Thermal vacuum chamber refurbishment and analysis. Master’s
thesis, California Polytechnic State University, San Luis Obispo, San Luis
Obispo, CA, June 2018.
134
APPENDICES
Appendix A
BAKEOUT PROCEDURE
This bakeout procedure is intended for the Blue TVAC located in at Cal Poly’s Space
Environments Lab 41-137. The purpose of a bakeout is to remove contaminants like
water vapor or other volatiles that were not removed with standard cleaning process
like IPA wiping. A bakeout is required when it is desired to reach the lowest vacuum
pressures possible. With a bakeout, ambient pressures in the chamber under 4.0e−6
Torr can be achieved at ambient temperatures.
1. Configure chamber for bakeout:
• Include all heater plates to maximize effectiveness
• Remove aluminum foil cover from door shroud and front of shroud, if
present
• Thoroughly clean all exposed chamber surfaces with IPA (Do not use paint
thinner or acetone on painted surfaces!)
• Include the cleaned test article for bakeout, if applicable, ensuring that
the maximum survival temperature of the article and melting points, va-
por pressures, and outgassing rates of the materials are known and not
exceeded.
• Verify the hot limit TC (23) is installed directly on the middle of one of
the exposed heater elements (this is extremely important!)
135
• Install the heater control TC (26) on the location where the set tempera-
ture is desired (usually the middle face of the middle heater plate, opposite
side as the element)
• Prepare LN2 if an extensive bakeout is desired
2. Rough pump to 5.0e−2 Torr and turn on the turbomolecular pump and then
ion gauge when the convectron gauge reads 0 per standard chamber procedures
3. After 1.0e−4 Torr is achieved, set a profile to ramp and soak at the desired
bakeout temperature at a slow ramp rate (¡0.5◦ C/min) on the F4 controller to
the, based on the test article maximum temperature. 90◦ C is the maximum
recommended control temperature to keep the hot limit from tripping. Switch
on the hot switch to activate the heater.
• Monitor the control temperatures and ion gauge pressure. If the pressure
climbs above 1.0e−3 Torr, turn down the heat until it stabilizes. If it is
approaching 1.0e−2 Torr with the heaters off, turn off the turbomolecular
pump and ion gauge to protect them from damage.
4. Hold at the bakeout temperature, noting the pressure change over time. Proceed
to the next step after 24 hours or if the pressure is below 7.5e−6 Torr while at
the bakeout temperature.
5. Switch off the heaters. If LN2 is available, flood the door shroud with LN2 by
setting set point 2 to -150◦ C on the F4 controller.
6. After the door shroud has achieved steady state temperatures and the pressure
has stabilized in the chamber, turn off the cooling and turn off the turbomolec-
ular and ion gauge.
7. After 40 minutes with the turbomolecular off, turn off the roughing pump and
vent the chamber.
136
8. After the chamber is vented, turn off the main breaker and open the chamber
door.
9. Using clean gloves and lint free wipes, use IPA to wipe down the door shroud
on the front face and back face with as much as possible by reaching through
the viewing port.
10. After the chamber is dry, close the door and proceed with regular testing. The
bakeout process can be repeated periodically as desired to remove adsorbed
water vapor or other contaminants.
137
Appendix B
UPDATED CHAMBER OPERATING PROCEDURE
The operating procedures are largely unchanged from their current state as doc-
umented in the manual, which was last updated on December 12th, 2019. This
appendix will discuss the relevant changes to each section as applicable.
B.1 Safety
No change. Note that the liquid nitrogen stored capacity, not mentioned in this
section, has increased from 35 liters to 180 liters. Nitrogen can displace oxygen in
the atmosphere, leading to hypoxia and death depending on the remaining oxygen
concentration. The existing oxygen sensor was unreliable when used in May 2021,
and likely needs re-calibration or replacement periodically, likely every year or sooner.
B.2 Chamber Overview
The interior image of the chamber is outdated and should be replaced with 6.1 from
this thesis.
B.3 Vacuum System
The system is now capable of pressures below 7.5e−6 Torr with bakeout of the chamber
and test article.
138
B.4 Heating System
The same except the platen is replaced by the three heater plates and the maximum
tested set temperature is 95◦ C, which is driven by a maximum coating temperature of
120◦ C near the heater strips, which is where the high limit TC is currently installed.
B.5 Cryogenic Cooling System
The same except the platen no longer part of the cooling system and the liquid
nitrogen tank has a relief pressure of approximately 22 psig instead of 7.25 psig.
The system was tested to -145◦ C in uncontrolled ramp testing, and to -80◦ C in
the plates thermal vacuum testing.
B.6 Operating Procedures
There are no changes to the operating procedures. Only the times to reach ¡50 mTorr
and times for the pressure to zero out, which are for information only, are changed
from 2 hours to 30 mins for ¡50mTorr and from 10 minutes to about 2 minutes for
the chamber pressure to zero out.
B.6.1 Starting The Cooling System
There are no changes to the procedures except for the ideal dewar pressure. The ideal
dewar pressure of between 4-7 psig is now 21-22 psig for the new liquid nitrogen tank.
139
B.6.2 During Normal Operation
As mentioned in the last subsection, the dewar pressure is now liquid nitrogen tank
pressure and the ideal range is 21-22 psig. There is unlikely to be a need to reduce
liquid nitrogen flow rates because of the larger tank and higher operating pressures.
B.6.3 Shutdown
These procedures are unchanged.
B.6.4 Opening the Dewar and Ordering LN2
The dewar is replaced by the liquid nitrogen tank and does not need to be opened or
vented by students at any time to fill. The filling is done by Praxair and they can be
called to come and refill the tank. Coordinate with them and Cody to ensure they
come to the right place to fill, use the same code or get the code from Cody. Delivery
and pickup is usually between 9:30am - 11:00am.
B.7 Contingencies
B.7.1 O2 Sensor Alarm (Evacuation Procedure)
These procedures are the same.
B.7.2 Increased Chamber Pressure
These procedures are the same.
140
B.7.3 Low Dewar Pressure (<3 PSIG)
This procedure is irrelevant because of the larger liquid nitrogen tank.
141
Appendix C
SOLIDWORKS FLOW SIMULATION THERMAL MODEL GUIDE
This guide is intended to be a basic demonstration of the various options and pa-
rameters relevant to the thermal model in this thesis that uses SolidWorks Flow
Simulation.
C.1 Adding a Test Article
Figure C.1 shows how to add a solid model to the existing thermal model assembly.
Figure C.1: Adding an test article part or assembly to the thermal model
Figure C.2 shows the options for moving the test article around in the assembly
as well as mating it to other parts like the platen or mounting rails.
142
Figure C.2: Moving and mating the test article in the assembly
C.2 Setting Up A Simulation
Figure C.3 shows how to load the SolidWorks Flow Simulation add on, as well as
view relevant fields in the flow simulation tab including different simulation cases as
well as material and surface properties, boundary conditions, and mesh setup.
Figure C.4 shows how to view parameters and how they apply.
Figure C.5 shows how change the heater plate temperature boundary condition.
This is what is changed to match the set point on the thermal vacuum chamber
controller. Other than defining the test article material and surface properties, this
is all that needs to be changed for hot case simulations with a new test article.
Figure C.6 shows how to change cases to the cold case and modify the temper-
ature boundary conditions for the cold case without flow simulation. For cold case
143
Figure C.3: Loading flow simulation and viewing cases and details
144
Figure C.4: Specific parameter selection
Figure C.5: Defining temperature boundary conditions for the hot case
145
simulations with flow, instead of temperature boundary conditions the user defines
flow boundary conditions in the same boundary conditions section.
Figure C.6: Defining temperature boundary conditions for the cold case,
without flow simulation
Figure C.7 shows how to add a radiative surface, and C.7 shows the process
for setting surface properties for individual surfaces of a test article. This process
is similar for defining material properties, initial conditions, and other boundary
conditions.
Figure C.9 shows how to define goals, which can be used to monitor parameters
during iteration as well as control convergence. Temperature goals for various surfaces
are particularly insightful for these thermal simulations.
C.3 Meshing, Solving, and Visualizing Results
Figure C.10 shows how to set up local and global meshes. The existing settings for
the meshes should be adequate however the level of detail can be easily adjusted.
146
Figure C.7: Adding a radiative surface
Figure C.8: Adding a radiative surface in detail
147
Figure C.9: Defining goals
Figure C.10: Basic meshing using global and local meshes
148
Figure C.11 shows how to run the mesher and solve the simulation.
Figure C.11: Running the mesher and solving the simulation
Figure C.12 shows how to view the mesh cut plot.
Figure C.13 shows how to visualize results relevant to this thermal model, includ-
ing probing locations for point temperatures and visualizing temperature distributions
across surfaces.
149
Figure C.12: Viewing the mesh
Figure C.13: Viewing temperature surface plots and probing results
150

Thermal Vacuum Chamber Modification Testing and Analysis

Uploaded by

Copyright:

Available Formats

Thermal Vacuum Chamber Modification Testing and Analysis

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Thermal Vacuum Chamber Modification Testing and Analysis

Uploaded by

Copyright:

Available Formats

THERMAL VACUUM CHAMBER MODIFICATION, TESTING, AND

the Faculty of California Polytechnic State University,

San Luis Obispo

of the Requirements for the Degree

Master of Science in Aerospace Engineering

TITLE: Thermal Vacuum Chamber Modification,

Testing, and Analysis

AUTHOR: Jared Lehmann

DATE SUBMITTED: July 2021

COMMITTEE CHAIR: Kira Abercromby, Ph.D.

Professor of Aerospace Engineering

COMMITTEE MEMBER: Pauline Faure, Ph.D.

Assistant Professor of Aerospace Engineering

COMMITTEE MEMBER: Jennifer Mott, Ph.D.

Assistant Professor of Mechanical Engineering

COMMITTEE MEMBER: Kim Shollenberger, Ph.D.

Professor of Mechanical Engineering

Thermal Vacuum Chamber Modification, Testing, and Analysis

can accommodate 6U Cubesats or larger for educational or research purposes.

nitrogen consumption, and modular components for flexibility in operation.

Analysis presented shows correlation between experimental results and a steady

demonstrate a maximum absolute difference in modeled vs experimental temperatures

Chamber performance is compared and characterized through a series of thermal

pressures under <7.5x10−6 Torr after bakeout procedures.

my life and my girlfriend, Brooke, who supported me through my thesis.

design, and my thesis.

1.1 Previous Capability vs ISO 19683 Requirements for TVAC Testing 12

2.1 Cooling sizing analysis summary table, best configurations high-

4.1 Empty test 1 thermal vacuum test profile . . . . . . . . . . . . . . . 77

5.1 Comparison of numerical model and experimental results for empty

1.1 Spacecraft Energy Exchange Diagram [17] . . . . . . . . . . . . . . 3

2.1 Concept shroud cooling tubing configurations . . . . . . . . . . . . 20

3.1 Tubing cross section after rolling process . . . . . . . . . . . . . . . 39

4.1 Methodology overview flowchart . . . . . . . . . . . . . . . . . . . . 67

5.1 Vent valve leak source outlined in red . . . . . . . . . . . . . . . . . 95

6.1 Chamber modifications (right) shown with previous configuration

manage the effects of exposure to these environments. The thermal environment

that these temperatures will be maintained throughout the mission. Mathematical

analysis is used to predict the performance of a spacecraft’s thermal control system

system meets design requirements.

A thermal vacuum (TVAC) chamber, used to conduct thermal vacuum testing,

capabilities of the TVAC chamber used.

previously modified MDA US systems thermal vacuum chamber located in California

and SolidWorks Flow Simulation.

1.2 The Thermal Space Environment

The thermal environment in space is dominated by radiative heat transfer because no

Figure 1.1: Spacecraft Energy Exchange Diagram [17]

heat selectively to the environment [17].

can be evaluated or only some specific objectives can be evaluated, depending on

individual component to a fully integrated satellite.

capabilities, and relevant specifications. For example, NASA’s General Environmental

Verification Standard (GEVS) requires prototype qualification test temperatures to

expected flight temperature [1].

cycling of hot and cold temperatures in an ambient or gaseous nitrogen environment,

the different ambient environment and difference in heat transfer modes.

which is important for radiofrequency payloads [6].

1.4 Thermal Modeling

Thermal modeling is any mathematical tool used to predict temperatures, typically

orbit and can be helpful to investigate anomalies or perform specialized operations or

analysis can take place.

q̇ = eq σ|Ti 4 − To 4 |[17] (1.3)