SHAPE MEMORY ALLOY ROBOTIC TRUSS 
 
 
 
 
 
 
Except where reference is made to the work of others, the work described in this 
 thesis is my own or was done in collaboration with my advisory committee.  
This thesis does not include proprietary or classified information. 
 
 
_________________________________ 
Lori Michelle Prothero 
 
 
 
 
 
 
 
 
 
Certificate of Approval: 
 
 
 
________________________________ ________________________________ 
Winfred A. Foster, Jr.               Robert S. Gross, Chair  
Professor      Associate Professor 
Aerospace Engineering   Aerospace Engineering 
 
 
 
________________________________ ________________________________ 
David A. Cicci               George T. Flowers 
Profesor     Dean 
Aerospace Engineering   Graduate School 
SHAPE MEMORY ALLOY ROBOTIC TRUSS 
 
 
 
 
 
Lori Michelle Prothero 
 
 
 
 
 
 
A Thesis 
Submitted to the 
Graduate Faculty of 
Auburn University 
 
in Partial Fulfillment of the 
Requirements for the 
Degree of  
Master of Science 
 
 
 
 
 
 
Auburn, Alabama 
August 9, 2008 
 iii
 
 
SHAPE MEMORY ALLOY ROBOTIC TRUSS 
 
 
 
 
Lori Michelle Prothero 
 
 
 
Permission is granted to Auburn University to make copies of this thesis at its 
discretion, upon request of individuals or institutions and at their expense.   
The author reserves all publications rights.  
 
 
 
 
 
 
________________________________ 
Signature of Author 
 
 
 
 
________________________________ 
Date of Graduation 
 
 
 
 iv
 
 
 
VITA 
 
 
 Lori Michelle Prothero, daughter of Karen R. S. Prothero and granddaughter of 
the late Myrtle R. Setzer and the late Robert L. Setzer, was born November 4, 1981, in 
Jacksonville, Florida.  She graduated as Valedictorian of Watauga High School in 
Boone, North Carolina, in June 2000.  In May of 2004, Prothero graduated Cum 
Laude and as a University Honors Scholar with a Bachelor of Aerospace Engineering 
degree from Auburn University in Auburn, Alabama.  During her first year at Auburn, 
Prothero began working in the Adaptive Aerostructures Laboratory for Dr. Ronald 
Barrett.  In her third year of enrollment, she received the Auburn University 
Undergraduate Competitive Research Fellowship to independently research and 
develop a space structure utilizing adaptive materials.  In April of 2004, Prothero 
presented her research findings at the National Conference of Undergraduate Research 
in Indianapolis, Indiana.  In December of 2003 she was accepted to be a Master's 
candidate at Auburn University starting in the fall of 2004 to continue her work with 
adaptive structures.  In the spring of 2006, Prothero's second generation prototype 
space truss using adaptive materials was tested in reduced gravity by the Auburn 
University NASA Reduced Gravity Student Team (Proposal ID: 2006-1767) in 
Houston, Texas aboard the NASA C-9 aircraft. 
 v
 
 
THESIS ABSTRACT 
SHAPE MEMORY ALLOY ROBOTIC TRUSS 
 
Lori Michelle Prothero 
Master of Aerospace Engineering, August 9, 2008 
(B.A.E., Auburn University, 2004) 
 
 
113 Typed Pages 
Directed by R. Steven Gross 
 
 
 The development, design, and analysis of a Shape Memory Alloy Robotic 
Truss (SMART) actuator is presented in this research paper.  SMART is a three-
degree of freedom actuator capable of large rotary and bending displacements using 
shape memory alloy (SMA) wires as the mechanism for actuation.  Using SMA 
actuator wires instead of conventional hydraulic actuators simplifies the overall 
complexity of design by reducing the number of working parts.  SMA actuator wires, 
because of vibrational dampening in the material itself, have a natural advantage over 
hydraulic actuators because they are not susceptible to large parasitic vibrations and 
long settling times inherent in hydraulic systems.   
With that said, the most radical development in the actuator design of SMART 
is that the entire structure acts as an actuator instead of actuation occurring at only a 
few synthetic joints.  That is to say that the amount of actuation is dependent on the 
 vi
length of SMART and corresponds directly to the structural stiffness of the truss.   For 
this reason, the truss backbone of the SMART actuator was designed to be structurally 
weak in torsion, strong in tension, and weak in axial bending to allow for the twisting 
and bending actuations.  The actuation force is provided by the contraction of SMA 
wires which are attached in a specific pattern, to be described in further detail later, to 
wire guides at nodes along the truss.  The force of the SMA wire?s contraction is 
distributed to the truss through the nodes at which the SMA is attached.  That is to say 
that the nodes connected to the SMA wire become closer, and as a result, the SMA?s 
contraction actuates the entire truss.   
The ability of the SMAs to contract is a unique material property of their 
crystalline structure to be trained at high heat to remember a desired length.  When 
cold, SMAs can be mechanically stretched easily; however, they immediately return to 
the remembered length when a heat stimulus is applied.  Exploiting this material 
property, electricity was supplied to one or multiple SMA wires in a simple circuit in 
which the SMA wires acted as the resistors.   The resistance produced heat in the SMA 
wires which then contracted in approximately a second to their remembered length.  
The result is that the entire truss actuated in a specified mode depending on which 
wires were heated.   
Results of SMART from ground based testing and reduced gravity testing 
aboard the NASA C-9 aircraft, while undergoing parabolic trajectories to simulate 
reduced gravity, demonstrated the feasibility of SMART as an actuator truss capable 
of large actuations and functionality in a reduced gravity environment such as space. 
 vii
 
 
ACKNOWLEDGEMENTS 
 
 The author would like to thank Dr. Ronald Barrett for being the primary 
faculty mentor throughout the research, for his expertise in shape memory alloy 
research, and for his awesome support of the project.  The author would also like to 
thank Dr. R. Steven Gross for assuming the position of major professor when Dr. 
Barrett accepted a position at the University of Kansas.  Special thanks are due Dr. 
Gross for his knowledge, support, and motivation throughout the writing process and 
without whom this thesis would not have been possible.  The author would like to 
thank the Auburn University Undergraduate Competitive Research Fellowship which 
helped fund this research.  The author wish to acknowledge the Auburn University 
NASA Reduced Gravity Student Flight Opportunities Project (RGSFOP) Team who 
flew with the research prototypes aboard the NASA C-9 research aircraft in March of 
2006. The author also wishes to thank the Lord to whom all praise is due for the 
ability to think and discover the awesome world He has created.   
 viii
 
 
Style manual or journal used: American Institute of Aeronautics and Astronautics 
Journal 
Computer software used:     Microsoft Word 2003, Microsoft Excel 2003               
 
 ix
 
 
TABLE OF CONTENTS 
 
LIST OF SYMBOLS ........................................................................................... xii 
LIST OF TABLES ............................................................................................. xiv 
LIST OF FIGURES ............................................................................................. xv 
I.  INTRODUCTION ............................................................................................. 1 
II.  SHAPE MEMORY ALLOY ........................................................................... 5 
2.1  CRYSTALLINE STRUCTURE ........................................................ 6 
2.2  TEMPERATURE EFFECTS ON SHAPE MEMORY ...................... 9 
2.3  PHYSICAL SIGNIFICANCE .......................................................... 10 
2.4  FLEXINOL? .................................................................................... 11 
III.  INITIAL DESIGN PHASE .......................................................................... 15 
3.1  STRUCTURAL DESIGN ................................................................ 15 
3.2  MATERIAL SELECTION ............................................................... 17 
3.3  KEVLAR? TRUSS MANUFACTURING ..................................... 18 
3.4  SMA WINDINGS ............................................................................ 21 
IV.  INITIAL DESIGN PHASE TESTING ........................................................ 22 
V.  INITIAL DESIGN PHASE DATA ............................................................... 29 
VI.  INITIAL DESIGN PHASE CONCLUSIONS.............................................. 32 
VII.  SECOND DESIGN PHASE ....................................................................... 34 
 x
7.1  NEW TRUSS STRUCTURE ........................................................... 34 
7.2  END CAP ASSEMBLY ................................................................... 36 
7.3.  BASE PLATE ASSEMBLY ........................................................... 38 
7.4  BI-DIRECTIONAL ACTUATION .................................................. 40 
VIII.  SECOND DESIGN PHASE TESTING .................................................... 41 
IX.  SECOND PHASE DATA ............................................................................ 43 
X.  SECOND PHASE CONCLUSIONS ............................................................. 45 
XI.  THIRD DESIGN PHASE ............................................................................ 46 
11.1  REDESIGN OF SMART CARRIER STRUCTURE ..................... 46 
11.2  DETAILED CONSTRUCTION INFORMATION ....................... 48 
11.3  CENTRAL SPINE .......................................................................... 49 
11.4  FIBERGLASS VERTEBRAE ........................................................ 49 
11.5  COLLETS ....................................................................................... 50 
11.6  END CAP ASSEMBLY ................................................................. 51 
11.7  ANCHOR SCREWS ...................................................................... 53 
11.8  END BASE ASSEMBLY .............................................................. 56 
11.9  TENSIONING SCREWS ............................................................... 58 
11.10  SHAPE MEMORY ALLOY (SMA) WIRES .............................. 59 
11.11  ELECTRICAL CIRCUIT ............................................................. 60 
XII.  GROUND AND REDUCED GRAVITY FINAL TESTING .................... 61 
XIII.  GROUND AND REDUCED GRAVITY TEST DATA ........................... 64 
13.1  BENDING SMART GROUND TEST ........................................... 64 
13.2  TWISTING SMART GROUND TEST ......................................... 66 
 xi
13.3  BENDING SMART REDUCED GRAVITY TEST ...................... 68 
XIV.  SMART PROGRAM CONCLUSIONS ................................................... 69 
REFERENCES .................................................................................................... 70  
APPENDIX A: PROPERTIES OF SHAPE MEMORY ALLOYS .................... 71 
APPENDIX B: PHOTOGRAPHS OF TEST SETUP  
                         FOR FIRST DESIGN ITERATION .......................................... 74 
APPENDIX C: PHOTOGRAPHS FROM SECOND  
                         DESIGN ITERATION ............................................................... 77 
APPENDIX D: PHOTOGRAPHS OF POSTTEST OF  
                         FINAL SMART TWISTING TRUSS ........................................ 81 
APPENDIX E: REDUCED GRAVITY TESTING OF   
                         SMART IN BENDING .............................................................. 83 
 
 
 xii
 
 
LIST OF SYMBOLS 
 
SYMBOL  DEFINITION      UNITS 
SMA   Shape Memory Alloy 
SMART  Shape Memory Alloy Robotic Truss 
NOL   Naval Ordnance Laboratory 
NiTiNOL  Nickel Titanium Alloy 
?   Displacement, Elongation, or Deflection   in./deg. 
?    Fraction of Phase Change from 0 to 1    none 
T   Temperature        ?C  
 
A
S
   Austenitic Phase Starting Temperature     ?C 
A
f
   Austenitic Phase Final Temperature      ?C 
M
S
   Martensitic Phase Starting Temperature     ?C 
M
f
   Martensitic Phase Final Temperature     ?C 
?   Registered Trade Mark 
?C   Degrees Celsius  
DC   Direct Curent        mV 
Tan
-1
   Inverse Tangent Function 
in. / ?   Inch 
? / deg   Degree 
 xiii
V   Electrical Voltage 
I   Electrical Current 
R   Electrical Resistance 
mV   Measure of Voltage, 10E-3 Volts 
mA   Measure of Amperage, 10E-3 Ampere 
sec   Unit of Time, Second      
g   Unit of Mass, Gram 
cm   Unit of Length, 10E-2 Meter 
lb   Unit of Weight, Pound 
?   Micro-, 10E-6 
 
?, ohm  Unit of Resistance 
 
psi   Unit of Pressure/Modulus Unit, Pound Per Square Inch 
GPa   Giga-Pascal, Unit of Stress (or Pressure), 10^9 Pascals  
%   Part of Whole, Percent 
cal   Unit of Heat, Calorie 
 xiv
 
 
LIST OF TABLES 
 
Table 2.4  NiTiNOL Flexinol? Published Technical ......................................... 11 
Table 5.1  Tabulated Data of Voltage ................................................................. 30 
Table 13.2  Twisting of SMART per SEGMENT ............................................... 67 
Table A.1  NiTiNOL Flexinol ? Published Technical Data ............................... 72 
Table A.2  Typical Properties of NiTiNOL SMA Alloys ................................... 73 
 
 
 xv
 
 
LIST OF FIGURES 
 
Figure 2.1.1  Austenitic (High Temperature) Crystal Structure ............................ 7 
Figure 2.1.2  Martensitic (Cold Temperature) Crystal Structure .......................... 7 
Figure 2.1.3  Deformed Martensitic (Cold Temperature) Crystal Structure ......... 8 
Figure 2.1.4  Diagram of Crystalline Transformations as  
                     SMA Heats and Cools, ? Denotes Elongation ................................. 8 
Figure 2.2.1  Phase Change Graph as a Function of Temperature ........................ 9 
Figure 2.3.1  Diagram of Shape Memory Alloy Changing  
                     Physical Shape and Phases ............................................................. 10 
Figure 2.4.2  Simple Test of SMA Material Characteristics ............................... 13 
Figure 2.4.3  Sample Test Data from Material Test of a 0.01?  
                     Diameter SMA Wire ....................................................................... 14 
Figure 3.1.1  Shape Memory Alloy Truss Schematic with Coordinate Axes ..... 16 
Figure 3.3.1  Schematic of a KEVLAR? Truss Lay-Up .................................... 19 
Figure 3.3.2  Enlargement of a Triangular Truss Section ................................... 20 
Figure 3.2.2  Triangular Truss Shown With Hole Indexing System ................... 20 
Figure 4.1.1  Photograph of Voltmeter, Ammeter, and DC Power Supply ......... 23 
Figure 4.1.2  SMA Truss on Test Stand with Torsional Relief Coil ................... 23 
Figure 4.1.3  Diagram of SMA Truss on Test Stand with Laser, Mirror,  
                      and Graph Board Measuring Setup ............................................... 24 
Figure 4.1.4  Diagram of Basic Geometry of Laser Reflection ........................... 25 
 
 xvi
Figure 4.1.5  Diagram of the Geometry Depicting Angular  
                                        Deflection of the Beam and Mirror ............................. 26 
Figure 4.1.6  Depiction of the Perceived Angle Change ..................................... 27 
Figure 5.1.1  SMA Truss on Test Stand After Testing ........................................ 29 
Figure 5.1.2  Graph of Voltage vs. Current ......................................................... 31 
Figure 5.1.3  Graph of Deflection vs. Voltage .................................................... 31 
Figure 7.1.1  Photograph of the post buckled Phase Two SMART .................... 35 
Figure 7.2.1  Photograph of SMA Loop with Washers for Anchoring ............... 36 
Figure 7.2.2  Photograph of SMA Loop and Washers Tightened ....................... 37 
Figure 7.2.3  Photgraph of the Finished End Cap ............................................... 38 
Figure 7.3.1  Close-up Photograph of the Base Plate Assembly ......................... 39 
Figure 9.1.1  Photograph of Near Maximum Deflection of the  
                     Second Prototype Twisting Actuator .............................................. 43 
Figure 11.1.1  Pattern for Fiberglass Vertebrae and  
                     Photograph of Finished Vertebra .................................................... 47 
Figure 11.3.1  Cap End of Central Spine ............................................................. 49 
Figure 11.4.1  Integrated Vertebrae on Central Spine ......................................... 50 
Figure 11.5.1  Collets against the Fiberglass Vertebrae ...................................... 51 
Figure 11.6.1  Photograph of the Tool-Hardened Steel Stub and Brass Tubing . 51 
Figure 11.6.2  Photograph of the Finished End Cap Assembly .......................... 52 
Figure 11.6.3  Photograph of End Cap Mating with End of Central Spine ......... 53 
Figure 11.7.1  Photograph of SMA Loop with Washers for Anchoring ............. 54 
Figure 11.7.2  Photograph of SMA Loop and Washers Tightened ..................... 54 
Figure 11.7.3  Anchor Screw with SMA Assembly Attached ............................ 55 
 xvii
Figure 11.7.4  Completed Anchor Screw Assembly ........................................... 56 
Figure 11.8.1  Base Assembly Before Ready for Tensioning Screws ................. 57 
Figure 11.8.2  Finished Base Assembly with Tensioning Screws ...................... 58 
Figure 11.10.1  SMA Wires on the Twisting Model of the SMART .................. 60 
Figure 12.1.1  Ground Based Test Setup of Bending Truss ................................ 62 
Figure 13.1.1  Photograph of SMART at Maximum Bending Deflection  
                       With and Without Displacement Measurements on Image .......... 65 
Figure 13.1.2  Side-by-Side Comparisons of Bending SMART Pretest,  
                       Posttest, and Posttest with Residual Bend Measurement ............. 66 
Figure 13.2.1  Photograph of Twisting Truss At Maximum  
                       Deflection Posttest ........................................................................ 67 
Figure 13.2.2  Plot of Twisting of SMART per Segment ................................... 67 
Figure B.1  Photograph of Test Setup Showing Laser,  
                   Test Stand, and DC Power Supply ................................................... 75 
Figure B.2  Enhanced photograph of the SMA Truss on the Test Stand ............ 75 
Figure B.3  Close-up Photograph of the Torsional Strain Relief Device ............ 76 
Figure B.4  Photograph of Test Setup with Test Stand, Laser, and  
                   Wall Graph Board ............................................................................. 76 
Figure C.1.  Frame of Maximum Twist Extracted from Video  
                    of Twist in Clockwise Direction ...................................................... 78 
Figure C.2  Frame of Truss Post Clockwise Twist with Relaxed SMA  
                   Extracted from Video of Twist in Clockwise Direction ................... 78 
Figure C.3.  Frame of Truss at Beginning of Counterclockwise Test Video ...... 79 
Figure C.4.  Frame Extracted Near Maximum Restoring Counterclockwise  
                    Twist from Test Video ..................................................................... 79 
Figure D.1.  Final Iteration of Twisting Smart Posttest with  
                    Measurements of Twist ................................................................... 81 
 xviii
Figure D.2  Close-up Photograph of Twisting SMART with Color Coding ....... 82 
Figure D.3  Researcher Lori Prothero with the Posttest SMART ....................... 82 
Figure E.1.  Auburn University ?Vomit Comet? Team at  
                    Team Readiness Review .................................................................. 85 
Figure E.2.  Ream Readiness Review Presentation with  
                   SMART Test Module at Right ......................................................... 85 
Figure E.3.  Demonstration of SMART Operations Inside Plexiglas  
                    Enclosure with NASA C-9 Aircraft in Background ........................ 86 
Figure E.4.  Unpacking and Presentation of Twisting SMART at  
                    Team Readiness Review .................................................................. 86 
Figure E.5.  Team Preflight Physiological Testing in Altitude Simulator .......... 87 
Figure E.6.  Andrew Wright during Physiological Test ...................................... 87 
Figure E.7.  Flight Crew and Flyers from March 9, 2006 at  
                    Johnson Space Center in Houston, Texas flight  
                    aboard the NASA C-9 Reduced Gravity Aircraft ............................ 88 
Figure E.8.  Ryan SMART Flyers Leureck and Vanessa Smith   
                    from March 9, 2006 Flight .............................................................. 88 
Figure E.9.  Andrew Wright, Backup Flyer for SMART Received  
                    Opportunity to Fly with Experiment from Another  
                    University on March 9, 2006 ........................................................... 89 
Figure E.10.  Ryan Leureck During Reduced Gravity Flight .............................. 89 
Figure E.11.  Andrew Wright During Reduced Gravity Flight ........................... 90 
Figure E.12.  Ryan Leureck and Vanessa Smith Operating  
                      SMART During Reduced Gravity Parabola .................................. 90 
Figure E.13.  Vanessa Smith and Ryan Leureck Observing  
                      SMART During Reduced Gravity Parabola .................................. 91 
Figure E.14.  Vanessa Smith and Ryan Leureck Observing  
                      SMART Through Top of Plexiglass Enclosure ............................. 91 
 
 xix
Figure E.15.  Flight Crew and Flyers from March 10, 2006 at  
                      Johnson Space Center in Houston, Texas flight  
                      aboard the NASA C-9 Reduced Gravity Aircraft .......................... 92 
Figure E.16.  Smart Flyers Michael Brennison and Meghan Brown  
                     with SMART Test Module on March 10, 2006 Flight ................... 92 
Figure E.17.  Smart Flyers Michael Brennison and Meghan Brown  
                     During Reduced Gravity Flight on March 10, 2006 ....................... 93 
Figure E.18.  Meghan Brown During Reduced Gravity ...................................... 93 
Figure E.19.  Michael Brennison Checking Computer In Flight ........................ 94 
Figure E.20.  Meghan Brown Checking SMART Bending Truss Experiment ... 94 
 
 1
 
 
I.  INTRODUCTION 
 
Shape memory alloys (SMAs) are a special class of metals with a unique 
crystalline structure which can be heat trained to remember a specific size and shape 
to which it returns when a heat stimulus is applied.  The unique characteristics of 
shape memory alloys (SMAs) were first discovered in the 1930s by Arne ?rlander 
while he was working with Cadmium-gold alloys.
1
  In 1962, William Buehler and 
other researchers at the Naval Ordnance Laboratory (NOL) discovered the memory 
effects in nickel-titanium alloy, which Buehler had previously renamed NiTiNOL in 
1956 to include the Naval Ordnance Laboratory acronym.
2
  Buehler discovered the 
ability of SMA to recover from high strains, of approximately 8% without plastic 
deformation.
2
  Eight percent strain recovery means SMA can stretch as much as 8% of 
the original length and return to its original length without any permanent deformation 
or change in the strength of the SMA.  Because of the ability to stretch and shrink 
repeatedly with strain recovery, SMA is the ideal material for mechanical devices 
which perform repetitive tasks.   
There are three general categories of SMA devices: devices which use the 
SMA shape recovery directly, stress recovery devices which exploit the stress of SMA 
when it is constrained, and actuators which use the force induced by SMA to cause 
movements in other objects.
3
   
 2
An actuator, as the name implies, is a mechanical device which causes 
something else to move.
4
  Actuators typically contain a large number of working parts 
such as gears and hydraulic pistons.  SMA actuators reduce the complexity of 
conventional actuators by replacing the numerous mechanical parts with a single wire 
capable of the same actuation force.
5
  Replacing hydraulic pistons and gears with a 
single wire also reduces the vibrational settling time, or shaking, as the actuator 
moves. Long vibrational settling times cause hydraulically mechanized machines to be 
more clumsy and imprecise in their movements.
4
  As a result, vibration is a major 
limiting factor in robotic design which is highly dependent on precise actuators 
designed for specific desired movements.   
There are already some SMA actuators for simple bending; however, only one 
of these designs by Padgett involves bending of an entire truss in two planes.
5
    
Howard, in his thesis on his design of a rotary SMA actuator, remarked that there exist 
very few designs for rotary actuators.
6
  Howard?s design is the only rotary design 
found of record which had more than one degree of freedom; however, both of its 
degrees of freedom are rotational and in a traditional joint-type actuator.
6
   Drawing 
from the success of a bending-truss actuator by Padgett and the need for rotary 
actuators, the Shape Memory Alloy Robotic Truss was conceived as a two-degree-of-
freedom bending actuator with a third rotary degree of freedom in the form of axial 
twisting of the truss for a total of three degrees of freedom.  
Because the design of SMART was so complex, the final design was 
accomplished in a number of iterative stages with smaller, more specific goals.  The 
first goal of this research project was to develop an axially twisting truss actuator as no 
 3
such device existed previously.  The original hypothesis was that helically wound 
SMA wires around the truss when heated, would cause the entire truss to twist.  After 
many design and prototyping iterations, a fully functional prototype twisting truss 
actuator was produced and tested.  The twisting truss actuator will be discussed in 
further detail later in this thesis.  After a twisting actuator had been developed, the 
second goal was to develop a bi-directional twisting actuator by incorporating 
opposing SMA actuator wires in a slightly modified truss.  The Bi-directional 
requirement meant that the truss had to be untwisting mechanically rather than untwist 
due to cooling wire relaxation and gravity.  After this second phase was accomplished, 
the truss had to be completely redesigned to accommodate bending.  Even though the 
third generation truss design radically differed from the first two designs, the method 
by which the truss actuated was still the same as the first and second generations with 
the added bending capability similar to Padgett?s actuator.  
The third generation truss reverted to a configuration first proposed by Padgett 
for his bending truss actuator which had a common central spine.  Along this spine 
were linearly spaced square trusses to which SMA attached linearly from truss to truss 
at each of the four corners; however, that is where the similarities between the Padgett 
design and SMART end.  There are many key differences between the second and 
third generation designs as well as between the third generation bending and twisting 
actuator and Padgett?s design.   
In the third design generation, two actuators were incorporated into the same 
truss structure: one was a twisting actuator and the other was a bending actuator.  By 
using the same truss for separate mode actuators, it was proposed that the truss could 
 4
be actuated as an integrated actuator of twisting and bending.  This was also supported 
by a test in the first phase of design in which a linear set of SMA wires was connected 
at the same time while the twisting mode SMA wires were activated and did not have 
any affect on the actuation.  In the third phase of design, small sliding insulators were 
added to the SMA wires in areas where the wires might have come into contact as a 
result of actuation of the bi-directional twisting.  The effective use of insulators proved 
that SMA wires in bending could also be isolated from twisting-mode SMA wires 
through insulators as well.  The fourth phase of the SMART project was extensive 
ground-based testing of both bending and twisting actuating trusses.  The fifth phase 
was the reduced gravity test of the SMART bending mode and twisting mode aboard 
the NASA C-9 reduced gravity aircraft.  The testing aboard the C-9 aircraft in 
parabolic freefall certified SMART and its structural components for reduced gravity 
operation such as found in space.   
Further research is needed to develop a truly-integrated bending and twisting 
truss actuator as well as an actuator suitable for operation in a space environment. 
 5
 
 
II.  SHAPE MEMORY ALLOY 
 
 Although Shape Memory Alloys were discovered much earlier by Arne 
Orlander, they have only been used in engineering applications since the discovery of 
NiTiNOL, nickel-titanium alloy, in the 1950's.
2
  William Buehler discovered the shape 
memory and high strain characteristics of NiTiNOL SMA quite by accident.  In 1959, 
Buehler and an assistant made six NiTiNOL bars in an arc-melting furnace.  Buehler 
dropped the first cooled bar on the concrete floor to see what the bar would do.  The 
bar made a thud indicative of some kind of vibrational damping within the material 
itself.  Buehler tested the other bars which were still warm and found the warm bars 
resonated with a ?bell-like quality sound.?  Buehler then cooled the warm bars in a 
nearby water fountain to see if the resonance of the warm bars was temperature 
dependent.  He found that the bars made the same thud sound as the cooled bar did.  
To test the reverse characteristic, he warmed the bars in boiling water, and the bars 
regained their resonance.   
Buehler continued to test the bars warm and cold and found the trend remained 
the same; the warm bars were resonant and the cold bars were damped.  Buehler 
understood that the change in the acoustic damping indicated that the atomic structure 
must change as a result of the temperature change.  The significance of the change in  
 
 6
atomic structure at different temperatures was not significant until the shape memory 
effects of NiTiNOL were discovered. 
 Raymond Wiley joined the Naval Ordnance Laboratory (NOL) in 1960 and 
worked on failure analysis of different metals including NiTiNOL.  In a briefing, 
Wiley demonstrated the fatigue resistance by repeatedly bending a NiTiNOL wire 
which had a diameter of one hundredth of an inch.  Repeated bending typically 
weakens metals until they eventually break; however the NiTiNOL wire did not break.  
While passing the wire around the boardroom table so that everyone could have a turn 
bending the metal, David Muzzey, one of the technical directors of the NOL, ?decided 
to see how it would behave under heat.?
 
  Muzzey held the accordion folded NiTiNOL 
strip in the flame of his pipe lighter and the wire became instantly straight by itself.  
Buehler, hearing about the incident, decided the shape recovery phenomenon was 
related to the earlier observation about sonic damping and resonance; however, the 
shape memory effects would be more useful.  Buehler?s earlier observation that the 
structure of the NiTiNOL?s atoms changed was later shown to be a change between 
two distinct crystalline structures depending on temperature. 
 
2.1  CRYSTALLINE STRUCTURE 
Shape memory alloys have two distinct phases called Austenite, which is the  
hot phase, and Martensite, which is the cold phase.
2
  The Austenitic phase is known 
for its regular lattice pattern as shown in the following figure.
7 
 
 
 
Figure 2.1.1  Austenitic (High Temperature) Crystal Structure.
2 
 
The darker, larger circles represent the Titanium metal ions and the lighter, smaller 
circles represent the Nickel metal ions. The pattern is repeated throughout the structure 
creating a regular face centered cubic crystal structure.
7
  The Martensitic phase occurs 
as the Austenitic phase cools and becomes distorted as in the diagram below.
2 
 
 
 
Figure 2.1.2  Martensitic (Cold Temperature) Crystal Structure.
2 
 
This type of alignment in which the crystals form parallelograms is called 
?monolithic.?
7
  The monolithic crystals align themselves as mirrored pairs called 
?twins.?  Twinning of the crystals does not affect the size or shape of the crystal 
significantly compared to the Austenitic lattice.  As a stress or displacement is applied 
to the lattice, the crystalline structure becomes detwinned; however, the original 
monolithic crystalline structure remains as in Figure 2.1.3. 
 7
 
 
Figure 2.1.3  Deformed Martensitic (Cold Temperature) Crystal Structure.
2 
 
The structure shown in Figure 2.1.3, which is still in the Martensitic phase, has been 
deformed.
2
  At this point heat can be applied and the lattice realigns in the Austenite 
form.  The diagram below shows the crystalline transformations as the structure heats 
and cools. 
 
 
AUSTENITE
MARTENSITE
(TWINNED)
MARTENSITE
(DEFORMED)
COOLING
HEAT 
RECOVERY
DEFORMING
?
 
 
 
 
 
 
 
 
 
Figure 2.1.4  Diagram of Crystalline Transformations as  
SMA Heats and Cools, ? Denotes Elongation.
2 
 
 
 8
2.2  TEMPERATURE EFFECTS ON SHAPE MEMORY 
 The shape memory effect of shape memory alloys is directly related to the 
change in temperature of the SMA.
8
  The figure below shows the change in phases as 
the metal cools or heats.  The symbol ? is the Martensite fraction and describes what 
state or fraction of phase transition the metal is in.  The value for ? is between zero 
0
1
?
T
M
f
A
f
M
s
A
s
Martensite
Austenite
Heating
Cooling
 
Figure 2.2.1  Phase Change Graph as a Function of Temperature.
4 
 
and one with zero denoting full Austenitic phase and one denoting full Martensitic 
phase. The symbol M
f 
 represents the temperature at which and below which the full 
Martensitic phase is attained.  The symbol M
s 
refers to the temperature at which the 
Martensitic phase begins.  The symbol A
s
 corresponds to the temperature at which the 
Austenitic phase begins, and A
f
 is the temperature at which full Austenitic phase is 
achieved.   
 
 
 9
2.3  PHYSICAL SIGNIFICANCE 
Recall in the Austenitic phase, the crystals align and remember a specific 
shape.  As the material loses heat, it changes phases to the Martensitic phase which 
retains the same shape since twinning does not change the exterior shape significantly.  
As more heat is lost, the Shape Memory Alloy (SMA) becomes fully Martensitic, and 
it is possible to physically realign the crystals of the SMA.  As the material is heated, 
it returns to the regular alignment of the Austenitic lattice without any trace of the 
Martensitic deformed shape.   
 
 
 
 
 
 
 
 
 
 10
 
Figure 2.3.1  Diagram of Shape Memory Alloy Changing Physical Shape and Phases.
2 
HEAT
HEAT COOL 
T = A
f 
 
 
 
T = M
f
 
 
 
T = M
f
 
 
 
 
T = A
f
 
 
 
 
 
 
T = M
f
 
 
Austenitic, Initial State 
 
 
 
 
Martensitic, Undeformed  
 
 
Martensitic, Mechanically 
Deformed 
 
 
 
Austenitic,  
Shape Recovered 
 
 
 
 
Martensitic,  Undeformed 
COOL 
 
  The physical significance of the phase change can be seen in Figure 2.3.1.  
The warm Austenitic phase metal retains its shape as it cools to the Martensitic phase.  
In the Martensitic phase, the metal is deformed mechanically.  As heat is applied, the 
metal returns to its original size and shape without any remaining deformation.   
2.4  FLEXINOL? 
 FLEXINOL? is the proprietary NiTiNOL alloy manufactured by Dynalloy, 
Incorporated and the particular NiTiNOL SMA wires used in SMART.
 9
  Dynalloy  
withholds the exact composition of FLEXINOL? as well as their processing and 
tempering of the alloy as trade secrets; however, they do supply extensive technical 
data specific to FLEXINOL? on their company website, www.dynalloy.com.  
According to the company website, this alloy was specifically engineered to be used in 
actuators; therefore, the company supplies the maximum pull force of a FLEXINOL? 
wire given its diameter.  It is also assumed that in the typical application, the SMA 
wires will be heated by electrical resistance in a circuit.  Dynalloy supplies the 
resistance of the wires in ohms per inch length as well as the current required in 
milliamps given the diameter of the wire in Table 2.4.
9
   
 
Table 2.4  NiTiNOL Flexinol? Published Technical Data
9
 
 
FLEXINOL? Technical Data 
Diameter 
Size 
(Inches) 
Resistance 
(Ohms per 
Inch) 
Maximum 
Pull Force 
(grams) 
Current 
at Room 
Temperature 
(mA) 
Contraction
Time 
(seconds) 
Off Time  
70? C Wire 
(seconds) 
Off Time  
90? C Wire 
(seconds) 
0.001 45 7 20 1 0.1 0.06 
0.0015 21 17 30 1 0.25 0.09 
0.002 12 35 50 1 0.3 0.1 
0.003 5 80 100 1 0.5 0.2 
0.004 3 150 180 1 0.8 0.4 
0.005 1.8 230 250 1 1.6 0.9 
0.006 1.3 330 400 1 2 1.2 
0.008 0.8 590 610 1 3.5 2.2 
0.01 0.5 930 1000 1 5.5 3.5 
0.012 0.33 1250 1750 1 8 6 
0.015 0.2 2000 2750 1 13 10 
0.02 0.12 3562 4000 1 18 15 
 11
 12
 
Using the resistance and current, it is easy to calculate the voltage across the 
wires given the circuit configuration.  In a simple, one strand circuit with power 
supply connected directly to the wire, the voltage obeys Ohm's Law as presented in 
equation (1) with voltage, V, current, I, and resistance, R. 
     V = I? R    (1) 
 According to the supplied technical data as well as sample testing in the 
laboratory, FLEXINOL? has been shown to have very good fatigue qualities with the 
laboratory tested capability of over a hundred-thousand flex and relax cycles without 
breakdown.
4
  FLEXINOL? is available in a range of gauges from 0.001 inch diameter 
to 0.02 inch diameter wires and has two different choices of transition temperatures: 
70?C and 90?C.
 9
  The initial smaller scale prototype SMART used 0.001 inch 
diameter SMA wires; however, the final prototypes which were almost ten times 
larger in size than the original SMART switched to 0.005 inch diameter SMA wires 
since there was heavier structure to actuate in the later designs.
 
The contraction time is approximately one second for each of the wires with 
the relaxation time varying between 0.1 seconds to a few seconds due to the diameter 
and the transition temperature of the chosen wire.
9
  The relaxation time varies because 
cooling is convective with ambient air and varies with surface area and thus with the 
diameter of the wire.
4
   Crystalline differences in the transition temperature also cause 
the cooling times to be inherently different.
4
   The technical data assumes that the 
wires are at room temperature and are isolated from radical air currents.
9
   
 
 The technical characteristics of FLEXINOL? samples were verified in the 
laboratory before selection for this project.  The experiment used on a range of 
different gauge wires consisted of a simple weight, below the maximum pull strength 
of each wire, hanging from a single strand of SMA wire as shown in Figure 2.4.2.   
 
 13
 
 
 
 
 
 
 
 
Figure 2.4.2  Simple Test of SMA Material Characteristics 
 
An electrical circuit was made by attaching leads to crimps ten inches apart on 
the SMA wire suspending the weight.  Note that the SMA did not carry the weight of 
the connecting lead wires; the bottom lead was draped across the test stand and the top 
was connected above the test section.  The leads connected the SMA to the positive 
and negative terminals of a variable voltage and variable current power supply with 
built-in ammeter and voltmeter.    
The voltage potential and current were slowly increased until the circuit was 
current limited, indicated by a red light on the power supply, at the desired test 
current.  The voltage was increased until a spike was observed and that voltage was 
recorded.  That voltage coincided with the moment at which the SMA lifted the 
weight.  Due to the lag time in the reaction of the SMA of about one second, the 
+
+
-
-
I 
V 
Voltag
e
Curren
t
I
0 
SMA Wire
Test Stand 
Weight 
Power Supply 
voltages were read directly from the power supply just as the ammeter registered the 
voltage spike instead of using only the visual observation of the wire contracting.  The 
test was repeated with other incremental currents tested and corresponding voltages 
recorded until the maximum current was exceeded and failure of the SMA occurred 
due to overheating.   
Figure 2.4.3 is the Voltage vs. Current test plot of a single strand of 0.01? 
diameter SMA.  The graph showed a clear linear relation with the slope of the linear 
regression giving the resistance of the circuit according Ohm?s Law.  The wire's 
theoretical resistance was calculated to be 5 ohms by multiplying the published 
resistance of 0.5 ohms per inch times the length of 10 inches.  The negligible increase 
of 0.1 ohms resistance of the actual circuit was due to the brass crimps, connecting 
wires, and resistance in the power supply with internal ammeter and voltmeter.  
Testing of different diameter wires with the same test circuit also had an increase of 
0.1 ohms which was determined to be a constant of the testing circuit. 
 
Voltage vs. Current
0.01" Diameter, 10" Long SMA
y = 5.1x
R
2
 = 1
0
1000
2000
3000
4000
5000
6000
7000
0 500 1000 1500
Current
Vo
l
t
a
g
e
 
 
 
 
 
 
 
 
 
Figure 2.4.3. Sample Test Data from Material Test of a 0.01? Diameter SMA Wire 
 14
 
15 
 
 
III.  INITIAL DESIGN PHASE 
 
The first challenge was to establish the initial and derived requirements of the 
design.  To create a twisting only actuator, the truss had to be structurally flexible in 
torsion while stiff in tension and compression.  The truss also had to accommodate 
connections and wire guides for the SMA to transfer force from the SMA as torque on 
the truss.  The SMA wires, which were heated by electricity, had to be attached in 
such a way that the wires would not touch each other or anything that would conduct 
electricity.  
 
3.1  STRUCTURAL DESIGN 
 The first requirement that the truss has to be flexible in torsion and stiff in 
tension and compression imposed that the truss be fixed in five of the six degrees of 
freedom: translation in the x-, y-, and z-direction and rotation about the x-, y-axis, 
while still being able to rotate easily about the z-axis as shown in Figure 3.1.1.   
 
Triangular trusses were used to constrain the xy-plane of the truss, which 
constrained translations in the x- and y-directions.  Stiff horizontal beam members 
connect the vertex of one triangular truss to the corresponding vertex of another as in 
Figure 3.1.1.  The arrangement of the three horizontal beams with a triangular base 
effectively constrained bending in the xz- and yz- planes as well as any linear 
 
translations along the z-axis.  Because the horizontal beams were perpendicular to the 
triangular trusses, this left the 3-D structure unconstrained in twisting about the z-axis 
since force carried in the horizontal beams were out of the xy-plane.   
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Triangular Members 
 Horizontal Members
y
z 
x 
 
Figure 3.1.1  Shape Memory Alloy Truss Schematic with Coordinate Axes. 
 
The joints between the triangular trusses and the horizontal beams had to 
withstand high stresses.  To constrain local rotations about the x- and y-axis, the joints 
between the triangle trusses and the horizontal members had to be reinforced with high 
shear strength epoxy.  The length of the horizontal members between triangular trusses 
also affected the strength of the truss.  Adding more triangular trusses and moving 
them closer together resulted in stiffening the structure but reduced twisting flexibility.  
Conversely, reducing the distance between triangular trusses and thus reducing the 
number of trusses along the horizontal beams decreased structural stiffness and 
increased twisting flexibility.  After several sample configurations were tested, the 
spacing between triangular trusses was sized appropriately to provide a stably 
16 
 
17 
stiffened structure which retained the a maximum amount of flexibility.  The 
triangular trusses which measure one inch on each side, were spaced every two inches 
along the horizontal beams. 
 
3.2  MATERIAL SELECTION 
Materials were carefully chosen for the truss members.  Boron filaments  
which were 0.050 inch diameter were selected for the horizontal members because of 
their high compressive strength.  Boron has a Young?s modulus of elasticity of 478 
GPa.
10
  Compared to the Young?s modulus of elasticity of steel, which is 200 GPa,
11 
boron is more than twice as stiff as steel.    
Strands of prepreg KEVLAR? which measured 0.04 inches by 0.012 inches 
were selected to make the triangular truss members.  KEVLAR? possesses very good 
tensile stiffness and strength properties while being poorer in compression.  Due to 
twisting actuation of the truss, the KEVLAR? trusses were only subjected to high 
tension loads such that given the scale of the application, the local stiffness of the resin 
in the KEVLAR? truss was sufficient for the nominal compressive loads.  The 
modulus of elasticity of KEVLAR? is 60 ? 120 GPa and the tensile strength is 3 GPa, 
or 3000 MPa.
12 
Incorporating the SMA actuation into the truss posed some difficulty.  To 
actuate the truss, the SMA needed to be attached to the truss to transfer the pull force 
evenly at each node.  The main obstacle was that SMA is difficult to bond with any 
material, including with itself, because the elasticity and strain recovery that make the 
material so useful, also prevents gluing, soldering, compression fitting, etc.   
 
18 
Many methods were tested before discovering that the most effective method for 
constraining SMA is to knot it with itself inside of a crimp.  The friction of the knot 
and crimp was sufficient to hold the SMA since the force required to pull the SMA out 
of it is usually more than the maximum load rated for the wire.  It was simply not 
practical to replicate the complicated knot/crimp joint while ensuring even pre-
tensioning between each crimped section of SMA at each of the three nodes of all the 
triangular trusses.  As a result, wire guides were incorporated into the KEVLAR? 
trusses through with the SMA could transfer the load as a contact load.  The SMA 
wires at each end of the truss were knotted and crimped with a crimp larger that the 
diameter of the wire guide so that it could not be pulled through the hole.   
Another concurrent complication in the design was that the SMA wires had to 
be electrically insulated from each other as well as from the boron filaments which are 
conductors.  Incorporating the wire guides into the triangular trusses was the logical 
solution since KEVLAR? is an electrical insulator.  SMA wire guides with a diameter 
of 0.01 inch were added at each node of the KEVLAR? triangle design with a 
geometric arrangement which prevented the SMA wires from touching each other or  
the boron.   
 
3.3  KEVLAR? TRUSS MANUFACTURING 
A single strand of KEVLAR? prepreg was tied in a pattern of knots, molded 
around piano wires larger than the diameter of the horizontal members, and pressed 
between two metal caul plates to create a triangular truss with holes large enough to 
accept the boron filaments at each vertex as well as the SMA as shown in Figure 3.3.1.   
 
 
 
 
 
 
 
 
 
 
 
Figure 3.3.1  Schematic of a KEVLAR? Truss Lay-Up 
 
Whole stacks of molded KEVLAR? using the top caul-plate surface as the 
bottom surface for the next mold were stacked together and clamped.  This method of 
efficient manufacturing was developed by Padget; however, adding knots and weaving 
the strands for added strength and adhesion of the epoxy was unique to this particular 
application.  The whole stack was placed in an autoclave to cure the impregnated resin 
in the KEVLAR? strand.  The KEVLAR? triangles were heated in the autoclave for 
eight hours at 170?C using standard composite materials manufacturing techniques.  
The result of the process can be seen in Figure 3.3.2 and Figure 3.3.3 which depict the 
regular pattern of the finished KEVLAR? triangular truss.  The wire guides were 
positioned symmetrically at each vertex of the triangular truss.  Figure 3.3.2 shows the 
indexing system of holes in the triangular truss sections.   
19 
 
 
 
20 
Boron Filament 
Interface 
Guides for  
SMA Wires 
 
 
 
 
 
 
 
 
 
 
Figure 3.3.2  Enlargement of a Triangular Truss Section to Show the Boron 
Filament Interface and Guides for the SMA Wires. 
 
 
 
2
B
B
R 
R 
S
S
L 
L 
1
S
B
R 
3 
L 
 
 
 
 
 
 
 
 
Figure 3.3.3  Triangular Truss Shown With Hole Indexing System. 
B Denotes the Boron Interface, L Denotes Local Left Upper Hole, R Denotes Local 
Right Upper Hole, S Denotes Shared Hole, and 1,2, and 3 Denote the Vertex. 
 
 
 
 
 
21 
3.4  SMA WINDINGS 
Three SMA wires were wound helically around the boron-KEVLAR? truss.  It 
is important to note that the strands were parallel so they never touched.  The 
following is a detailed description of the SMA windings using the indexing system.  
The first SMA wire started at vertex 1 in the hole L of truss one.  Then the 
same SMA wire passed through the second truss at hole S of vertex 2.  Then that SMA 
wire passed through hole L at vertex 3 of the third truss and so on down the length of 
the truss.  Two other SMA wires started separately at vertex 2 and vertex 3 of truss 1 
and continued in a pattern similar to that of the first wire.  None of the wires crossed 
each other or came into contact with the boron.  Three opposing SMA wires were 
similarly strung starting at the other end of the truss.  The wires ran through the shared 
and right holes of the KEVLAR? triangles.  The end result was that three SMA wires 
ran in a clockwise direction three wires ran in a counterclockwise direction without 
any of the wires crossing.   
Opposing pulling wires heated alternately act much like bicep and tricep 
muscles do to actuate and restore position of an elbow joint.  As one muscle (or SMA 
wire) contracts, the opposing muscle (or SMA wire) extends and vice versa for motion 
in the opposite direction.  For the laboratory test, however, only twisting in only the 
clockwise direction was tested to prove the basic concept of SMA torsional actuation.  
The counterclockwise winding would work in an identical manner for actuation in the 
opposite direction; therefore, testing of the counterclockwise windings was redundant.  
The redundant wires were left in the truss to show that the opposing wires do not come 
in contact with the other set of SMA wires. 
22 
 
 
IV. INITIAL DESIGN PHASE TESTING 
 
 
A truss constructed of four triangular KEVLAR? trusses secured with 
cyanoacryllate at two inch intervals on three boron filaments as previously described 
was placed on a test stand.  Small sections of brass tubing were crimped at each end of 
pre-stretched SMA wires to provide a surface to pull against the end of the truss and 
transfer the force to the truss.  Contact wires which had a 0.15 inch diameter were then 
soldered to the brass crimps.  A simple electrical circuit was made with a variable DC 
power supply, external ammeter used to accurately measure the current, internal 
voltmeter used to accurately measure the voltage, and SMA wires, which acted as 
three resistors in parallel.  The electrical resistance in the SMA wires provided the heat 
to the SMA.  The electrical supply/control system is shown in Figure 4.1.1.   
The truss was attached to the test stand with tape around the top contact wires 
as in Figure 4.1.2.  The straight contact wires naturally resist torsion and would have 
restrained the actuation of the SMA Truss.  Therefore, the contact wires at the bottom 
of the truss were connected to a torsional strain relief device.  The torsional strain 
relief device was simply a length of the same type of wire used as contact wires which 
had been coiled into a 0.4 inch diameter.  The coiled shape allowed the SMA Truss to 
deflect more easily.   
 
 
 
 
 
 
 
 
 
 
Figure 4.1.1  Photograph of Voltmeter, Ammeter, and DC Power Supply. 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 4.1.2  SMA Truss on Test Stand with Torsional Relief Coil at Bottom 
23 
A laser mirror was mounted just above the torsional strain relief.  A laser was 
focused onto the mirror and the reflection from the mirror onto a graph board to 
measure the deflection as in Figure 4.1.3.   
 
LASER
Test Stand
Graph Board
 
 
 
 
 
 
 
 
 
 
Figure 4.1.3  Diagram of SMA Truss on Test Stand with Laser, Mirror, 
and Graph Board Measuring Setup. 
 
The focal lengths of the laser and graph board were set such that one inch 
deflection on the graph board corresponded to one degree of deflection of the truss.  
The geometry of the reflected laser beam from the deflection mirror is shown in 
Figure 4.1.4. 
 
 
 
 
24 
Opaque 
Wall
Flat Mirror
Laser
Theoretical 
Origin of
Incident Ray
Reflected 
Ray
Normal Line
?
r
?
i
Fixed Distance
 
Figure 4.1.4  Diagram of Basic Geometry of Laser Reflection 
 
 In Figure 4.1.4, an opaque wall was positioned behind the laser source at a 
fixed distance parallel to a flat mirror.  The symbol ?
i
 represents the angle of incidence 
between the source laser beam and the normal line to the mirror.  The angle of 
reflection ?
r
 is the angle between the normal line to the mirror and the reflected laser 
beam from the mirror.  The angle of incidence and the angle of reflection are equal 
according to the Law of Reflection presented in Equation (2). 
??? ??
ir
     (2) 
Figure 4.1.5 depicts the same setup as the flat mirror deflects an angle ?. 
The perceived angle of deflection on the opaque wall is ?.   
25 
Opaque 
Wall
Flat Mirror
Laser
Theoretical 
Origin of
Incident Ray
Original Reflected Ray
Original Normal Line
Fixed Distance
?
?
New Normal Line
New Reflected Ray
?
 
Figure 4.1.5  Diagram of the Geometry Depicting Angular 
Deflection of the Beam and Mirror 
 
 
Since the position of the origin beam stayed fixed relative to the wall, the 
perceived angle change ? is really the sum of the change in angle of the incident beam 
and of the reflected beam from the previous mirror position when the mirror was 
parallel to the wall. 
ri
??? ?+?=    (3) 
Defining the change in angle as the difference between the new angle and old 
angle and using the identity from Equation (2) results in Equation (4): 
0
??? ?=?     (4) 
Equation (3) then becomes: 
 ?? ?= 2     (5) 
Recalling that the mirror deflection is ? and geometric relations for 
perpendicular lines, the deflection of the normal line is also ?.  The angle ? is also the 
26 
true angle of deflection and the change in deflection of the angle of incidence and 
angle of reflection.  Therefore, Equation (5) can be rewritten in terms of ?. 
?? 2=     (6) 
Since linear distances are more easily measured than angles, angle ? is 
calculated geometrically from the linear displacement of the reflected beam on the 
opaque wall.  Using a fixed distance between the opaque wall and the mirror which is 
many orders of magnitude larger than the linear displacement, the wall may be 
assumed to be approximately perpendicular to the original reflected beam.  Then the 
calculation of ? becomes simple trigonometry as depicted in Figure 4.1.6. 
Figure 4.1.6 Depiction of the Perceived Angle Change ? with Wall Assumption 
Opaque 
Wall 
New Reflected Ray 
Flat 
Mirror 
y
? 
x 
Original 
Reflected Ray 
 
The definition for the tangent function of an angle in a right triangle is used to 
determine angle ?.  The equation for ? is: 
y
x
Tan =)(?     (7) 
Solving the equation for y, the previous equation becomes: 
)(?Tan
x
y =     (8) 
27 
Realizing that for calculation purposes it would be much easier if ? of 1? corresponds 
to x = 1?, we let ? = 2? recalling from Equation (6) that ? is twice ?.  Plugging in 
Equation (8), we get: 
"65.28
)1(
"1
?
?
=
Tan
y    (9) 
Thus, the experiment used a fixed distance between the mirror and the wall of 28.65? 
so that an inch of lateral deflection of the laser beam on the wall corresponded to a 1? 
twist of the truss.   
A range of voltages was applied to the circuit, and the current as well as the 
angular deflection were measured and recorded.  The data was afterward compiled 
into a table and plotted.  More photographs from the test setup may be found in the 
Appendix. 
28 
 
 
V. INITIAL DESIGN PHASE DATA 
 
The twisting of the SMA truss can be seen qualitatively below in the 
photograph.  Also presented is a tabulated chart of the data collected in Table 5.1. 
 
Figure  5.1.1  SMA Truss on Test Stand After Testing. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
29 
30 
Table 5.1  Tabulated Data of Voltage With Output Current and Displacement 
Voltage Current Displacemet (?) Voltage Current Displacemet 
(?) 
(mV) (mA) (inch) (mV) (mA) (inch) 
121 10 0.165 2540 200 -0.276 
227 18 0.365 2660 210 -0.843 
382 30 0.360 2660 210 -0.692 
504 40 0.375 2790 220 -1.184 
635 51 0.455 2790 220 -0.572 
750 60 0.400 2930 230 -1.543 
902 72 0.402 2930 230 -0.956 
992 79 0.403 3050 240 -1.882 
1130 90 0.385 3050 240 -1.324 
1260 100 0.308 3180 250 -3.550 
1260 100 0.412 3180 250 -2.083 
1380 110 0.302 3170 260 -11.6 
1380 110 0.380 3170 260 -11.3 
1500 120 0.268 3210 260 -11.6 
1500 120 0.371 3210 260 -11.3 
1640 131 0.244 3360 270 -12.3 
1640 131 0.319 3360 270 -11.4 
1770 141 0.152 3400 270 -12.3 
1770 141 0.283 3400 270 -11.4 
1880 150 0.068 3440 280 -13.9 
1880 150 0.194 3440 280 -13.4 
2010 160 -0.110 3480 280 -13.9 
2010 160 0.137 3480 280 -13.4 
2150 170 -0.091 3530 290 -23.4 
2150 170 0.089 3530 290 -23.0 
2280 180 -0.176 3560 290 -23.4 
2280 180 -0.161 3560 290 -23.0 
2410 190 -0.464 3610 300 -23.9 
2410 190 -0.261 3610 300 -23.1 
2540 200 -0.589 3650 300 -23.9 
 
 
A graph of the Voltage vs. Current of the twisting actuator is provided in 
Figure 5.1.2, and a plot of Voltage vs. Deflection of the actuator is presented in Figure 
5.1.3.   
 
 
 
Figure  5.1.2 Plot of Voltage vs. Current 
Voltage vs. Current
0
500
1000
1500
2000
2500
3000
3500
4000
0 50 100 150 200 250 300 350
Current (mA)
V
o
l
t
ag
e (
m
V
)
 
 
Figure  5.1.3  Plot of Deflection vs. Voltage 
  
Voltage vs. Deflection
-30
-25
-20
-15
-10
-5
0
5
0 50 100 150 200 250 300 350
Voltage (mV)
D
e
fl
ec
ti
o
n
 (d
eg
r
ees
)
Deflection vs. Voltage
Voltage (mV)
31 
 
32 
 
 
VI.  INITIAL DESIGN PHASE CONCLUSIONS 
 
The proof of design SMA Truss actuator performed well during testing and 
showed that high angular SMA actuation of a truss section is feasible.  When heat was 
applied to the SMA, the SMA caused the truss to twist as is evident qualitatively in 
Figure 5.1.1.  The plot of Voltage vs. Current, Figure 5.1.2, shows a linear relationship 
between Voltage and Current.  Recalling Ohm?s law that voltage is current times 
resistance, the slope of that graph represents the resistance of the system.  This is 
consistent with the published resistance for the type of SMA wire used in this 
experiment.  The maximum angular deflection was approximately twenty-three 
degrees as shown in the graph of Voltage vs. Deflection, Figure 5.1.3.   
SMART met the design requirement that the entire truss twist to accommodate 
rotational actuation and was an overall success.  Since this stage of the design was an 
intermediate proof of design feature to a more complex design, further testing and 
development of this particular actuator were abandoned to pursue the integrated 
actuator.  Future recommendations for developing this particular intermediate actuator 
as its own functional design would include integrating the opposing SMA wires for bi-
modal twist and untwist.  Also, extensive testing should be performed to determine the 
fatigue life of the actuator after repeated cycles of twisting.  A mathematical model of 
the twisting actuation would also need to be developed using numerical analysis 
 
33 
methods as the data did not readily converge to any of the 300 closed form solutions 
which were attempted by using curve fit software.  The technological strides and proof 
of concept discoveries made in the initial design phase directly translated into the 
success of the final design of the Shape Memory Alloy Robotic Truss. 
 
 
34 
 
 
VII.  SECOND DESIGN PHASE 
 
The initial design phase of the SMA Truss actuator demonstrated the capability 
of a rotationally actuated truss using helically wound SMA actuator wires.  The next 
set of design requirements required the capability of bi-directional actuation, so a new 
actuator was build which could accommodate bi-directional actuation.  
 
7.1  NEW TRUSS STRUCTURE 
Since the small scale of the initial design had made construction tedious and 
time consuming, the overall scale was increased to facilitate faster prototyping.  The 
new triangular KEVLAR? trusses had exactly the same thickness and orientation of 
wire guides as the initial design; however the distance between vertices of the new 
triangles was increased from one inch to two inches and the wire guide hole diameter 
was increased to accommodate new horizontal members and new 0.005 inch diameter 
SMA actuator wires.  The boron filament horizontal members of the initial design 
were replaced with 0.05 inch diameter graphite rods due to the limited selection of 
boron diameters on hand and the hazard of working with boron filaments which are 
prone to shatter into sharp splinters.   
A working prototype truss of 14 inches tall with KEVLAR? triangles fixed  
with cyanoacryllate at two-inch intervals was constructed and attached firmly to a base 
 
of two sheets of 0.16 inch thick modeler?s plywood laminated with cyanoacryllate.  
An end cap was made on the other end of the truss with two laminated pieces of the 
same modeler?s plywood.  Both the base and end cap were bored approximately 0.1 
inches deep to accept the ends of the graphite rods at right angles.  The graphite rods 
were firmly attached in their holes in the base and end cap with cyanoacryllate. 
Since the brass crimping technique in the initial design was difficult to adjust 
to assure proper pre-tensioning of the SMA wires, a new wire termination system was 
developed.  The new system works much like guitar tuning pegs at the base and 
anchor screws at the end cap.   
Once the wires were properly pretensioned, it was observed that the 
KEVLAR? trusses buckled under the force of the tightened SMA wires.  Plywood 
triangle webbing was added to the KEVLAR? triangles to carry the compressive load 
because the trusses buckled during pretensioning of the SMA.   
 
 
 
 
 
 
 
 
 
Figure 7.1.1  Photograph of the post buckled Phase Two SMART 
35 
 
7.2  END CAP ASSEMBLY 
 The end cap assembly consisted of a wooden end cap with anchor screws.  
Two holes were bored through the end cap at each vertex: one set of holes inside the 
triangle made by the graphite rods at the vertices and the other set outside that triangle.  
The six holes had a diameter just larger than a 0-80 size screw and were drilled at an 
offset angle so they would be accessible to a mini wrench.   
 The anchor screws were 0-80 size, stainless steel (silver in color), 0.5? long 
screws with hex heads and a number of #0 stainless steel washers and 0-80 steel nuts.  
The following is a step by step description of an anchor screw assembly:   
1. A piece of 0.005 inch diameter SMA wire was threaded straight through a 
first size #0 stainless steel washer and a second washer of the same type. 
2. The SMA wire was then looped over and threaded in the opposite direction 
back through the first washer while skipping over the second washer as 
shown in the following photograph.   
 
 
 
 
 
 
 
 
Figure 7.2.1  Photograph of SMA Loop with Washers for Anchoring 
 
 
 
36 
 
3. Then the wire was pulled snug against the washers as shown in the 
following photograph. 
 
 
 
 
 
 
 
 
 
 
Figure 7.2.2  Photograph of SMA Loop and Washers Tightened 
 
4. The two washers were pressed such that the aligned one on top of the other 
and a 0-80 size stainless steel screw was pushed through the washers as in 
Figure 7.2.3. 
5. A 0-80 size nut was added to the screw and tightened such that the washers 
were firmly pressed to the hexagonal head of the screw and to each other. 
This kept the SMA from being pulled through the washers. 
6. Another #0 washer was added to the screw.  The screw was inserted into 
the end cap and another washer was added on the screw at the other side of 
the end cap.   
7. Then one wire of a three-wire cable was stripped of the insulation and 
wound around the screw on top of the last washer with a 0-80 stainless 
steel nut tightened firmly down on top of the wire to secure it.  Note that no 
bare wire was left outside of the interface with washer and nut.  Refer to 
Figure 7.2.3. 
37 
 
8. Steps 1-7 were repeated for each of the remaining five anchor screws 
paying special attention to which wire was attached to which screw 
obeying the following:  Anchor screws on the inside of the triangle were 
attached to one to each wire of a three-wire cable and the outer anchor 
screws were attached one to each wire of a separate three-wire cable.  
Refer to section 7.4 for an explanation of the wiring of the SMA circuit. 
Below is the completed end cap.  See Figure 7.2.3. 
 
 
 
 
 
 
 
 
Figure 7.2.3  Photgraph of the Finished End Cap  
 
7.3.  BASE PLATE ASSEMBLY 
 The base plate assembly consisted of the base plate, tensioning screws and 
wooden feet glued to the base plate which lifted the base plate so that the assembly did 
not sit on the screws.  A wooden cover plate attached with rubber bands to the base 
plate insured that the ends of the tensioning screws did not come into contact with 
anything.  See Figure 7.3.1.   
 
 
38 
 
 
 
 
 
 
 
 
Figure 7.3.1  Close-up Photograph of the Base Plate Assembly 
 
 The tensioning screws were 2-56 size, steel alloy (black in color), 0.5? long 
screws with hex end caps to be used with an Allen wrench.  The following is a 
description of the tensioning screw assembly: 
1. A 2-56 sized steel nut followed by two #2 steel washers, another 2-56 steel 
nut, and a final #2 washer were threaded onto a 2-56 steel alloy screw.   
2. The screw assembly was put through a hole in the base plate and another 
nut was added to the reverse side but not tightened to the base. 
3. To attach the SMA, the first two nuts which sandwiched the washers were 
unscrewed slightly and the SMA was laid between the first two washers.   
4. The SMA was wrapped around the screw about four times in the same 
direction as the helical winding it made with the truss. 
5. The two nuts on either side of washers were tightened and the SMA was 
firmly attached while the screw was free to rotate in its hole in the base. 
6. The whole screw was rotated to tension the SMA and then the anchoring 
nut on the underside was tightened to the base. 
39 
 
40 
7. Then one wire of a three-wire cable was stripped of the insulation and 
wound around the bottom end of the last nut with another 2-56 steel nut 
tightened firmly down on top of the wire to secure it.  Note that no bare 
wire was left outside of the interface of nut and nut.  Refer to section 7.4 
for an explanation of the wiring of the SMA circuit. 
 
7.4  BI-DIRECTIONAL ACTUATION 
Both clockwise and counterclockwise windings of 0.005 inch diameter SMA 
actuator wires were attached to the wire guides of the KEVLAR? trusses.  Two 
separate circuits were made: one with the three clockwise wound SMA wires and the 
other with the three counterclockwise wound SMA wires.  A different tri-wire cable 
was connected to each set of clockwise or counterclockwise wires at each end.  Tri-
wire cables which had a green wire in them were connected to the clockwise wound 
SMAs, and tri-wire cables which had a red wire connected to the counterclockwise 
wound SMAs.  Red insulated alligator-type connecting clips were soldered to each of 
the cables coming off of the end cap while black alligator-type connectors were 
soldered to the cables from the base cap.   
The black alligator connector of the green (clockwise SMA) cable was 
connected to the negative terminal of the power supply.  The red alligator connector of 
the green cable was attached to the positive terminal of the power supply.  This made a 
circuit of three resisting SMA wires in parallel as in the initial design.  Refer to the 
section on the initial design for specific circuit setup.   
 
 
41 
 
 
VIII.  SECOND DESIGN PHASE TESTING 
 
The second design of the Shape Memory Alloy Robotic Truss was tested to 
show twisting actuation and opposing twisting actuation of the same truss.  Two 
separate circuits were made.  The set of clockwise SMA wires wound around the truss 
were each attached at the anchor/cap end screws to one strand of a tri-wire cable with 
a distinguishable green strand.  The set of counterclockwise SMA wires wound around 
the truss were likewise attached individually to a strand in the tri-wire cable with a red 
strand.  The green strand cable from the cap end of the truss (with a soldered red 
alligator clip) attached mechanically to the positive terminal of a variable voltage 
power supply.  The other green strand cable from the base end of the truss (with a 
soldered black alligator clip) attached mechanically to the negative terminal of the 
variable voltage power supply.  This created a parallel circuit of three SMA wires 
acting as resistors.  An ammeter internal to the power supply was connected to the 
circuit in series and measured the input current in the circuit.  A voltmeter internal to 
the power supply was connected in parallel to the circuit to measure the voltage drop 
of the entire circuit.  The second circuit was made similarly by replacing the 
corresponding alligator clips of the counterclockwise circuit to the same power supply 
with internal ammeter and voltmeter.   
 
42 
Current was supplied to one circuit at a time to determine the maximum 
twisting displacement in either direction.  When a circuit was connected, current and 
voltage were turned up in small increments from zero until the ammeter registered and 
motion was observed.  Leaving the power supply set to this current and voltage, the 
circuits were alternated between clockwise and counter clockwise SMA sets.  Then to 
determine the maximum actuation, the current and voltage were increased by about 
20% to increase the speed of actuation and to determine the maximum automated 
actuation.  The circuits were then tested alternately.  Finally the circuits were 
disconnected and the truss was manually twisted to determine the true absolute 
maximum actuation of the truss itself.  The experiment was videotaped and 
photographed against a background graph board at approximately six inches behind 
the truss to determine the maximum deflections.   
 
 
 
IX.  SECOND PHASE DATA 
 An image from the twist experimental is presented below.  Bi-directional 
actuation was achieved through alternate activation of the clockwise and 
counterclockwise SMA wire sets.  Though the actuation was difficult to measure; it is 
evident from the video captured and in still images that the truss was capable of at 
least 60 degrees of twist with a slight preference to the clockwise direction.   
 
 
 
 
 
 
 
 
 
 
Figure 9.1.1  Photograph of Near Maximum Deflection of the  
Second Prototype Twisting Actuator 
 
 It was also discovered during testing that there are twisting limits.  As the truss 
twists more, it becomes increasingly difficult to twist the truss further.  At higher 
angles of twist in the truss, the graphite rods which are relatively stiff to twisting 
43 
 
44 
locally in themselves, must twist locally to allow the entire truss to twist.  This can be 
observed in the previous test image.  At approximately 180 degrees of rotation, there 
is also an interference problem with the graphite rods which point touch each other.  
These conclusions of twist limitations were conducted by manual twisting of the truss 
structure. 
 
 
45 
 
 
X.  SECOND PHASE CONCLUSIONS 
 The goal of bi-directional actuation was achieved in this design using opposing 
helical windings of SMA wires.  This actuator prototype performed similarly to the 
previous truss; however, there were some notable conclusions which fell out of the 
testing.  The sizing and stiffness of the rods and fixed ends of the truss affected the 
amount of actuation achieved.  The original triangular trusses of prepreg KEVLAR? 
were inadequately stiff locally to prevent buckling.  This required a redesign of the 
triangles during testing to add a more a rigid plywood web.  Redesign of the SMA 
carrier triangles was noted as an opportunity for improvement in future prototyping.  
During testing, it was shown that increasing the voltage and current above minimum 
operational activation would increase the speed of the actuation; however, it would 
probably lead to shorter fatigue life.  Also, an important observation was made that the 
more truss module segments in the truss, the more twist actuation is capable because 
the amount of twist contributed by each truss cell is additive.  Because there was 
discovered a limitation on the amount of twist which can be achieved in this design, a 
redesign of the support structure was recommended as essential to achieve better 
twisting capabilities. 
 
 
 
 
46 
 
 
XI.  THIRD DESIGN PHASE 
 The third design iteration marked a complete metamorphosis of the design with 
the resulting third iteration resembling the earlier designs very little; however, the 
essential technology behind the twisting actuation had changed very little.  The third 
design phase also added bending actuation capability to the twisting truss.   
 
11.1  REDESIGN OF SMART CARRIER STRUCTURE 
 New consideration was give to the earlier Padget design which had utilized a 
central graphite rod as a spine with equally spaced KEVLAR? squares with integrated 
wire guides at the corners and cross members through the center to make a hole 
through which the spine was placed and secured with cyanoacryllate.  The third 
iteration of SMART utilized Padget?s heritage application for a bending actuator with 
SMA at four corners.  Applying the lessons learned from the previous two SMART 
iterations that KEVLAR? are not sufficiently stiff in compression for more than the 
smallest scales, the wire guide structure was completely redesigned in a different 
material.  Prepreg fiberglass fabric was chosen as the new material for the SMA wire 
carriers because of its increased compressive strength as well as for being an electrical 
insulator.   
Instead of using a graphite spine along which the square SMA carriers were 
permanently affixed with cyanoacryllate, a steel rod of approximately 0.60 inches was 
 
used.  Further design influences came from nature and from snake vertebrae which 
have a great deal of articulation due to the allowed rotation of their vertebrae in the 
spinal column.  Therefore, the fiberglass SMA carriers were allowed to rotate freely 
about the central steel spine and were only constrained from translating along the 
spine by the addition of collets on either side of the fiberglass vertebrae.  When the 
first prototype fiberglass vertebrae was produced, it was discovered that it did not slide 
well about the steel spine because of the ?brooming? delaminations which form as 
holes are drilled; therefore steel washers just larger than the diameter of the steel spine 
were added as inclusions in the layup to serve as crude bearings.  There were no 
observed delaminations around the modified vertebrae with steel washers for bosses as 
is evident in the photograph of a finished vertebra.  The steel washer bosses also 
served as alignments for the templates to be affixed to the cured blanks.   
 
Figure 11.1.1  Pattern for Fiberglass Vertebrae, left  
and Photograph of Finished Vertebra, right 
 
Further design alterations were made to accommodate six holes for SMA wires 
in each node instead of the original five.  It was thought that by tying the SMA wires 
to each vertebra that the force would be better transferred to the vertebra instead of 
relying on contact with the rough drilled hole edge.  Also, this method was developed 
47 
 
48 
to maximize the pull force of each unit of the spine and allow each vertebra to work 
equally.  This concept was quickly abandoned when it was discovered that the wires 
degraded rapidly from tying in such tight knots and the heating in the wires was not 
uniform.  The wires located at the ends were doing the most pulling.  Rather than 
remake the vertebrae for only three holes at each corner (one for bending wire, 
clockwise twist, and counterclockwise twist), the vertebrae were used as is with 
leaving three of the holes empty.  The pull force of the wire was allowed to 
equivalence itself along the entire length of the spine, and it was discovered that 
bearing directly on the rough hole guides did not affect the wires significantly.  It 
should be noted as a risk that frequent chafing from the holes could cause premature 
wear on the SMA wires.  The inclusion of some sort of silicone inserts or grommets 
would alleviate this chafing.   
 
11.2  DETAILED CONSTRUCTION INFORMATION 
Two SMARTs were made almost identically except that one was wired for 
bending and the other for twisting.  Both bending and twisting model SMARTs use 
the same base components as follows:  
A. Central Spine 
B. Fiberglass Vertebrae 
C. Collets 
D. End Cap Assembly 
E. Anchor Screws 
F. Base Assembly 
G. Tensioning Screws 
H. Shape Memory Alloy (SMA) Wires 
I. Electrical wiring 
 
 
 
11.3  CENTRAL SPINE 
The central spine is a steel rod with a diameter between 0.058 and 0.065 
inches.  One end of the steel rod is inserted into the base assembly and secured with 
cyanoacryalate.  See Section on Base Assembly.  The other end of the rod is sharpened 
to a point and mates with the End Cap.  Refer to section on End Cap Assembly.   See 
photograph below. 
 
 
Figure 11.3.1  Cap End of Central Spine 
 
 
11.4  FIBERGLASS VERTEBRAE 
 ?Clover-shaped? Fiberglass Vertebrae are located every three inches along the 
Central Spine and simply supported by collets on either side.  The individual vertebrae 
are made from four plies of epoxy pre-impregnated fiberglass with a washer placed 
middle layer to reinforce the central hole in the finished vertebrae.  The blank lay-up is 
placed between two caul plates and heated in an autoclave for two hours at 177?C.  
Patterns are placed on the blank to align with the washers.  Then the pattern is cut out 
using a band saw, and the holes are drilled with a hand drill.  The holes accommodate 
49 
 
the Shape Memory Alloy wires and act as electrical insulators between the Central 
Spine and each SMA wire at the nodes.  The Vertebrae slides onto the steel spine and 
is secured between two Collets which prevent the vertebrae from translating.  See 
section 11.5 on collets. 
 
Figure 11.4.1  Integrated Vertebrae on Central Spine 
 
11.5  COLLETS 
 The Collets are appropriately sized to fit over the steel Central Spine and 
contain set screws which tighten them onto the steel spine.  The set screws will either 
require a jeweler?s screwdriver or appropriate sized hex key/Allen key.  Collets may 
be purchased at a hobby supply store.  Each Fiberglass Vertebrae is sandwiched 
between two collets to prevent translation up or down the Central Spine.  Please note 
that the Collets do not prevent rotation. 
 
 
 
 
 
 
 
 
 
50 
 
 
Figure 11.5.1  Collets against the Fiberglass Vertebrae 
 
 
11.6  END CAP ASSEMBLY 
 The End Cap is made up of four components:  brass tube stock, tool-hardened 
steel stub, cap circle, and Anchor Screws.  Anchor Screws will be covered in the next 
section.  Brass tubing which just fits over the steel spine is cut to approximately two 
inches in length.  A piece of hardened steel stub which is less than one inch long, 
conveniently acquired from old drill bits, is chosen with a diameter such that it fits 
snugly inside the brass tubing.  The steel is bonded inside the brass with 
cyanoacryalate. 
 
Figure 11.6.1  Photograph of the tool-hardened steel stub and brass tubing 
 
 
51 
 
 Next, a piece of ?light-ply? hobbiest?s plywood with about a 1/8 inch thickness 
is cut into a disc with a diameter of approximately two inches.  A depression is made 
into the center of the disc to accept the brass/steel stub assembly.  The brass/steel 
assembly is bonded with cyanoacryalate to the End Cap.  A filet of cyanoacryalate is 
made at the joint to make the connection much stronger.  A filet may be made by 
applying a cyanoacryalate accelerating spray and cyanoacryalate epoxy in alternating 
layers to build up a filet.  Holes are drilled for the Anchor Screws according to the 
same pattern as the Fiberglass Vertebrae; however, only a single hole is drilled per 
pair in the pattern. 
 
Figure 11.6.2  Photograph of the Finished End Cap Assembly 
 
 Graphite lubricating dust, commonly used to lubricate locks and is available in 
any hardware store, is put into the brass tubing to allow smooth rotation when the End 
Cap Assembly is mounted to the sharpened end of the steel spine.  Please note that 
because the Steel Spine only contacts the End Cap Assembly at the sharpened point 
52 
 
and graphite lubrication reduces friction, the End Cap Assembly acts like a frictionless 
bearing and is free to rotate. 
 
Figure 11.6.3  Photograph of End Cap Mating with End of Central Spine 
 
Please note, the first bending model of S.M.A.R.T. does not have a brass tubing/steel 
stub on the End Cap Assembly.  The end of the Steel Spine simply fit into a hole 
drilled halfway though the End Cap.  This constrains the rotational degree of freedom 
at the End Cap; however, the bending actuation of the S.M.A.R.T. does not create any 
torsion on the end.  As a result, rotation at the end cap is insignificant for the bending 
model, and the basic End Cap is sufficient for the Bending S.M.A.R.T.   
 
11.7  ANCHOR SCREWS 
 The Anchor Screws are the attachment nodes for the Shape Memory Alloy 
(SMA) wires at the End Cap Assembly.  The Anchor Screws are 0-80 steel alloy 
(black in color) screws that are 1 inch in length and have a hex cap head for use with 
53 
 
an Allen key.  The following instructions are the step by step description for installing 
the Anchor Screws: 
1. Take one 0-80 steel alloy (black in color) screw that is 1 inch in length. 
2. Thread one 0-80 stainless nut until it is snug against the cap head of the screw.  
(The original Bender model omitted this step) 
3. Take two size #0 stainless steel washers and thread a piece of 0.005 inch 
diameter SMA wire through both of them.   
4. Then loop the SMA wire around the first washer and go though the first 
washer like in the photograph below: 
 
 
 
 
Figure 11.7.1  Photograph of SMA Loop with Washers for Anchoring 
 
5. Pull the SMA snugly against the washers.  See Photograph Below. 
 
Figure 11.7.2  Photograph of SMA Loop and Washers Tightened 
54 
 
6. Put the 0-80 alloy screw with washer through the washer such that the head of 
the screw ends up on the same side as the SMA wire ends.  See Figure 10. 
7. Push washers up the screw until snug against the nut.  See Photograph Below. 
 
 
Figure 11.7.3  Anchor Screw with SMA Assembly Attached 
 
8. Thread another 0-80 stainless nut on the screw until snug against the washers 
so that the washers are sandwiched between nuts. 
9. Thread another 0-80 stainless nut on the screw and stop before reaching the 
other nuts/washers leaving approximately a 1/8 inch gap for Bending model 
S.M.A.R.T., a 1/8 inch gap for Twisting model S.M.A.R.T. outside wires, or a ? inch 
gap for the Twisting model S.M.A.R.T. inner wires. 
10. Add a size #0 stainless steel washer against the last nut on the screw. 
11. Put screw assembly into hole on End Cap and push tight against cap.   
12. Add a size #0 stainless steel washer against the outside facing side of the End 
Cap. 
13. Add a 0-80 stainless nut until it is snug against washer and End Cap.   
 
55 
 
 
Figure 11.7.4  Completed Anchor Screw Assembly 
 
14. Check to see that the SMA wire is firmly attached to the screw assembly and 
that the screw assembly is firmly attached to the End Cap. 
15. For electrical contacts, it may be desirable to add another 0-80 nut to the screw 
with a 1/8 inch gap.  Then add an electrical lead, as pictured above, and secure with 
another nut.  This step may be modified accordingly or omitted as in the case of 
alligator type electrical connectors. 
 
11.8  BASE ASSEMBLY 
 The Base Assembly consists of a base block which is about 4 inches long with 
embedded 2-56 plated nuts for 2-56 steel alloy (black in color) Tension Screws and a 
Mounting Base Plate.  The end of the Central Spine attaches to the Base Assembly in a 
hole drilled into the top of the block.  The joint is secured using cyanoacryalate in the 
hole and a filet of cyanoacryalate around the outside of the joint.  The filet is made 
56 
 
similarly to the filet around the End Cap.  A hole is drilled into the base block which is 
the diameter of the 2-56 screws at a depth of approximately ? inch.  Then a punch is 
used to make a depression in the wood centered on the screw hole.  The nut is placed 
in the depression and then pounded into the wood until it is flush with the surface of 
the base block.  Cyanoacryalate is applied around the edge of the nut to assure the nut 
is firmly attached.  The process is repeated on each of the four sides of the base as 
shown in the following picture.   
 
Figure 11.8.1  Base Assembly Before Ready for Tensioning Screws 
 
57 
 
 
Figure 11.8.2  Finished Base Assembly with Tensioning Screws 
 
11.9  TENSIONING SCREWS 
 The Tensioning Screws work similar to tuning pegs on a guitar which enable 
tightening and loosening of wires using a mechanical screw.  The Tensioning Screws 
are made from 2-56 steel alloy (black in color) 1 inch screws.   
 On the Bending Model, as pictured above in Figure 13, two nuts are screwed 
together onto the screw with only a slight gap between them.  The screw assembly is 
then turned partially into one of the nuts mounted in the Base Assembly (See 
Photograph Above).  The SMA wire is wound in the gap in a clockwise motion so that 
as the screw is tightened into the Base Assembly, the wire gets tighter.  As the screw is 
loosened, the wire becomes loose.  The SMA must wrap around the screw several  
 
58 
 
59 
times and the nuts must be tightened together for the SMA wire to be secured  
by friction.   
 The Twisting Model works similarly to the Bending Model with the addition 
of two washers between the nuts which give more surface area to clamp onto the  
SMA wire. 
 Tension should be relatively close in each of the wires and may be felt by 
touching the wires.  The wires should not have any slack but should not be so tight 
that the S.M.A.R.T. is almost rigid. 
 
11.10  SHAPE MEMORY ALLOY (SMA) WIRES 
 Nickel Titanium Shape Memory Alloy Wires with a transition temperature of 
70?C and a diameter of 0.005 inches are used.  Wires should be cut to approximately 
42 inches to insure overrun.  Please refer to directions about Anchor screws and 
Tensioning Screws for instructions on how to connect the wires.   
 On the Bending Model, wires run linearly from one node to the next in the 
same corresponding hole.  On the Twisting Model, sets of wires run clockwise and 
counterclockwise as in Figure 11.10.1.  The inner set of wires use Teflon tubing 
inserted over the wires in the stringing process.  The tubing is approximately 2/3 of  
the length of the exposed wire length nodes. 
 
 
60 
 
Figure 11.10.1  SMA Wires on the Twisting Model of the SMART 
 
11.11  ELECTRICAL CIRCUIT 
 Power is supplied to the SMA wires by a manual power supply.  Tandem 
pulling SMA wires are connected in parallel.  This means that pairs of wires on the 
Bending model are wires in parallel, and the four outer wires on the Twisting model 
are wired in parallel separately from the four inner wires on the Twisting model which 
are wired together in parallel.   
 Power Requirements for the S.M.A.R.T.?s are as follows: For the Bending 
Model, a current not exceeding 1 amp is supplied at a voltage of between 30 and 35 
volts.  For the Twisting Model, a current not exceeding 1 amp is supplied at a voltage 
of around 30 volts.   
 
 
61 
 
 
XII.  GROUND AND REDUCED GRAVITY FINAL TESTING 
 The ground based tests of both the twisting and bending were conducted using 
nearly identical setups as the previous iterations of SMART.  The actuation modes 
were tested separately on identical SMARTs with one wired with helical SMA for 
twisting and the other wired linearly for bending.  This was done to accommodate the 
need for redundant test articles for two days of testing aboard the NASA C-9 aircraft.  
If one SMART failed, the other would be ready to go that day and be available for 
reworking to accommodate the other mode testing during the night between flights.  A 
third truss structure was made; however, it was not wired and was there in case spare 
parts were needed or could be wired in case both trusses failed or the flown truss could 
not be reworked.   
 In ground based testing, SMART was suspended by its base which was 
clamped in a fixture.  A graph board was placed just behind SMART so that 
displacements could be measured from still and video images taken of the truss.  
Additionally, a mirror was mounted to the tip of the truss and a laser beam bounced 
off of the mirror onto a graph board to measure deflections of the tip of SMART more 
accurately.   
 
 
 
 
 
 
 
 
 
 
 
 
Figure 12.1.1  Ground Based Test Setup of Bending Truss 
 
The phase one design test procedure in Section IV has more details regarding laser 
angle/displacement calibration.  One inch deflection of the laser beam corresponded to 
an upward bend of the truss of approximately one degree of bend in the truss.  The 
activation current of both SMARTs were found experimentally as before in previous 
testing procedures and photographs were taken of the max deflections so that the 
defections in the images could be measured.  A suspended yardstick served as the 
plumb line measurement for the images, and a horizontal scale at the bottom was used 
to measure tip directions in the horizontal plane.   
 The reduced gravity testing aboard the NASA C-9 aircraft tested the bending 
actuated SMART both days of flight due to the twisting truss failing in pre-flight 
check out and lack of time between flights to rework/rewire the bending to twisting or 
the spare.  The activation voltage/current was found experimentally in pre-flight 
checkout.  SMART was suspended from the top of a Plexiglas enclosure and a mirror 
62 
 
63 
was mounted on the tip of SMART to reflect the laser beam onto a graph board.  
Smart was manually turned on just before the 30 seconds of parabolic path freefall for 
approximately 60 cycles of parabolas per flight.  Digital video equipment continuously 
captured movements of the reflected laser point as SMART actuated and stored the 
video to the hard drive of a laptop in the test enclosure.  The enclosure was mounted 
on four large dampers so that vibrations from the C-9 aircraft would have less effect 
on the data.  The detailed circuit, test setup, photographs from the flights, as well as 
the operations manual for SMART for the reduced gravity testing by the Auburn 
University NASA Reduced Gravity Student Team is available in the Appendix.   
 
64 
 
 
XIII.  GROUND AND REDUCED GRAVITY TEST DATA 
 The two SMARTs were tested in ground based tests.  Both SMARTs were 
slated to fly aboard the NASA C-9 aircraft in reduced gravity; however, due to 
technical difficulty, the Twisting SMART was scratched and the Bending SMART 
was flown on both days of flight on March 9-10, 2006 at NASA Johnson in Houston, 
Texas.  The members of the Auburn University team who performed the in flight 
operations of SMART were Michael Brennison, Ryan Leureck, Megan Brown, and 
Vanessa Smith, Andrew Wright and Christopher Worley served as ground crew and 
backup crew for flight operations.   
 
13.1  BENDING SMART GROUND TEST 
 In testing the bending truss, the laser beam left the graph board before max 
deflection.  Also, because of the high deflection of the tip of SMART, the laser beam 
no longer focused on the mirror.  The still image of bending SMART at maximum 
actuation shows a tip deflection of 13.5 inches in the horizontal and a tangent angle of 
the tip section of 58 degrees from the vertical yellow yardstick.  Measurements were 
calculated from image processing of the photograph taken at maximum deflection 
from a camera mounted perpendicular to test plane.   
 
 
 5? 
 13.5? 
58? 
Figure 13.1.1  Photograph of SMART at Maximum Bending Deflection With and 
Without Displacement Measurements on Image 
 
 The base of SMART was observed to rotate in its fixture.  Using the same 
photograph captured at maximum bending deflection, the angle deflected by the base 
was measured.  The 5 ? deflection of the base was added to the tip deflection of 58? 
which resulted in a total truss deflection of 63? in bending.   
Following the bending test, there was an observable post-test residual 
deformation of the truss.  The residual tip deflection was 1.5 inches from pretest and a 
bending angle residual of 7?.  Repeating the test consistently yielded the same 
deflection and residual post deflection.   
 
65 
 
 
7? 
Figure 13.1.2  Side-by-Side Comparisons of Bending SMART Pretest, Posttest, and 
Posttest with Residual Bend Measurement 
 
13.2  TWISTING SMART GROUND TEST 
 The ground based testing of the twisting truss used a mirror mounted on the tip 
of the truss in the same setup as the bending test.  No appreciable measurements were 
captured from the laser point as the point only reflected the local twisting of the end 
cap.  Local twisting occurred all along the length of the truss, and as a result, the total 
deflection of the entire structure was calculated from the measured twisting of each 
segment visible at maximum actuation of SMART. 
 
 
 
 
66 
 
 
 
 
 
Figure 13.2.1  Photograph of Twisting Truss At Maximum Deflection Posttest.  
(Photograph is Rotated 90 Degrees Counterclockwise) 
 
Table 13.2  Twisting of SMART per SEGMENT 
Segment Node From Node To Deflection 
1 Base Vertebra 1 0? 
2 Vertebra 1 Vertebra 2 75? 
3 Vertebra 2 Vertebra 3 15? 
4 Vertebra 3 Vertebra 4 45? 
5 Vertebra 4 Vertebra 5 45? 
6 Vertebra 5 Vertebra 6 30? 
7 Vertebra 6 Vertebra 7 90? 
8 Vertebra 7 Vertebra 8 75? 
9 Vertebra 8 Vertebra 9 90? 
10 Vertebra 9 Vertebra 10 90? 
11 Vertebra 10 Vertebra 11 90? 
12 Vertebra 11 End Cap 90? 
TOTAL TWISTING DEFLECTION: 735? 
 
Figure 13.2.2  Plot of Twisting of SMART per Segment 
Twisting of SMART Per Segment
0
10
20
30
40
50
60
70
80
90
100
01234567891011213
Segment Number
D
e
f
l
e
c
t
io
n
 in
 
D
e
g
r
ee
s
 
 
 
 
 
 
 
67 
 
68 
13.3  BENDING SMART REDUCED GRAVITY TEST 
 The Bending SMART flew aboard the NASA C-9 Aircraft and operated safely.  
Post flight analysis of the structure certified that the structure remained intact for the 
entire flight and still maintained mechanical functionality following reduced gravity 
operations.  Observations from the flight test crew indicated that SMART worked as 
anticipated and experienced bending deflections of more than 90 degrees during 
reduced gravity sets.  The video captured in-flight did not have an adequate frame of 
reference to measure true deflection of SMART from normal.  What can de derived 
from the film is that there was a gain of approximately 12 degrees in the max 
deflection during the reduced gravity sets when compared to the max deflection in 
normal gravity.  This was measured from the deflection of the laser beam on a 
graphboard between max deflection at constant gravity and max deflection at reduced 
gravity.   
 
 
 
69 
 
 
XIV.  SMART PROGRAM CONCLUSIONS 
 There were many strides as well as many lessons learned over the course of 
SMART development, and the final SMART design benefits from both.  SMART 
successfully satisfied the fit, form, and function of a degree of deflection bending and 
twisting actuator.  Ground based testing produced approximately 70 degrees bending 
and a little over 2 full rotations of the structural components.  Further reduced gravity 
testing of SMART in bending resulted in additional bending by conservative estimates 
of more than 10 degrees.  Successful flight tests aboard the NASA C-9 aircraft 
certified SMART as a reduced gravity actuator and raised the design?s technology 
readiness level one step closer to space operations and the transition from prototype to 
production. 
 Further design refinements to be considered in the future are: to find a better 
method of electrical insulation for SMA wires that will not hinder even heating of the 
wires, work out actuation control to achieve intermediate actuation, integrating 
operations of bending and twisting in tandem on a single SMART, creating a 
mathematical model to further explore the additive effects of segment actuation of a 
chain of segments, and studying out-gassing of materials in SMART and considering 
using materials which would be more suitable in a vacuum and under very hot and 
very cold temperatures.    
 
 
70 
 
 
REFERENCES 
 
1.  Fromer, J., Sectionally Heated Shape Memory Alloy Actuator for Multiple Position 
Aircraft Roll Control, M.S. Thesis, Auburn University, Auburn, Alabama, 2002. 
 
2.  Srinivasan, A. V., and McFarland, D. M., Smart Structures: Analysis and Design, 
Cambridge University Press, Cambridge, United Kingdom, 2001, pp. 26-72 
 
3.  Schwartz, M., Encyclopedia of Smart Materials, John Wiley & Sons, Inc., New 
York, New York, Vol.2, 2002, p. 923. 
 
4.  Barrett, R. M., Introduction to Adaptive Aerostructures, Short Course Presented at 
Auburn University, May 2003. 
 
5.  Howard, N. A.  Design, Construction, and Analysis of a 2-Degree-Of-Freedom 
Shape   Memory Alloy Rotary Actuator, M.S. Thesis, Auburn University, Auburn, 
Alabama, 2000. 
 
6.  Padgett, D. A.  The Design and Construction of a Shape Memory Alloy Tail Boom 
Actuator.  41st Aerospace Sciences Meeting and Exhibit. 6-9 January 2003, Reno, 
Nevada 
 
7.  Otsuka, K. and Wayman, C. Shape Memory Materials, Cambridge University 
Press, Cambridge, United Kingdom, 1999. 
 
8.  Banks, H., Smith, R., and Wang, Y., Smart Material Structures: Modeling, 
Estimation, and Control, John Wiley & Sons, Inc., New York, New York, 1996, pp. 
18-26. 
 
9.  Dynalloy Co. Web Page, http://www.sma-inc.com/Actuator.html, 2004. 
 
10.  Kelly, A., Concise Encyclopedia of Composite Materials, MIT Press, Cambridge, 
Massachusetts, 1989, pp 33-35. 
 
11.  Hibbler, R., Mechanics of Materials, Prentice Hall, Upper Saddle River, New 
Jersey, 2000, p. 92. 
 
71 
 
 
APPENDIX A 
PROPERTIES OF SHAPE MEMORY ALLOYS 
 
72 
Table A.1  NiTiNOL Flexinol ? Published Technical Data
5 
 
Flexinol? Technical Data 
Diameter 
Size 
(Inches) 
Resistance 
(Ohms/Inch) 
Maximum 
Pull 
Force 
(grams) 
Approximate* 
Current at 
Room 
Temperature 
(mA) 
Contraction* 
Time 
(seconds) 
Off Time 
70? C 
Wire** 
(seconds) 
Off Time 
90? C 
Wire** 
(seconds) 
0.001 45 7 20 1 0.1 0.06 
0.0015 21 17 30 1 0.25 0.09 
0.002 12 35 50 1 0.3 0.1 
0.003 5 80 100 1 0.5 0.2 
0.004 3 150 180 1 0.8 0.4 
0.005 1.8 230 250 1 1.6 0.9 
0.006 1.3 330 400 1 2 1.2 
0.008 0.8 590 610 1 3.5 2.2 
0.01 0.5 930 1000 1 5.5 3.5 
0.012 0.33 1250 1750 1 8 6 
0.015 0.2 2000 2750 1 13 10 
0.02 0.12 3562 4000 1 18 15 
 
 
73 
Table A.2  Typical Properties of NiTiNOL SMA Alloys
2 
 
Typical Properties of NiTiNOL Alloys 
Physical Properties 
Melting Point  1300 ?C 
Density 6.45 g/cm^3 
  0.233 lb/in^3 
Electrical Resistivity: 
     Austenite ~100 ??cm 
     Martensite ~70 ??cm 
Thermal Conductivity:     
     Austenite 0.18 W/cm?C 
     Martensite 0.085 W/cm?C 
Similar to 3000 Series 
Corrosion Resistance 
Stainless Steel or Ti 
Alloys 
Mechanical Properties 
Young's Modulus:     
     Austenite ~12E6 psi 
     Martensite ~4.6E6 psi 
Yield Strength:     
     Austenite 28-100E3 psi 
     Martensite 10-20E3 psi 
Ultimate Strength 130E3 psi 
Elongation Failure 20-30 % 
Transformation Properties 
Transformation 
Temperature -100 to 200 ?C 
Latent Heat of 
Transformation 40 cal/g atom 
Shape Memory Strain 8.4% Maximum 
 
 
 
74 
 
 
APPENDIX B 
PHOTOGRAPHS OF TEST SETUP  
FOR FIRST DESIGN ITERATION 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure B.1  Photograph of Test Setup Showing Laser,  
Test Stand, and DC Power Supply 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure B.2  Enhanced photograph of the SMA Truss on the Test Stand.  The Dark 
Lines Trace the Boron Filaments.  The Lighter Gray Lines trace the SMA Wires 
 
75 
 
76 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure B.3  Close-up Photograph of the Torsional Strain Relief Device  
Which is the White Coil in the Photograph 
 
 
 
Figure B.4  Photograph of Test Setup with Test Stand, Laser, and Wall Graph Board 
 
 
77 
 
 
APPENDIX C 
PHOTOGRAPHS FROM SECOND DESIGN ITERATION 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure C.1.  Frame of Maximum Twist Extracted from  
Video of Twist in Clockwise Direction. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure C.2  Frame of Truss Post Clockwise Twist with Relaxed SMA Extracted from 
Video of Twist in Clockwise Direction.  Note the Residual Deformation. 
78 
 
79 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure C.3.  Frame of Truss at Beginning of Counterclockwise Test Video.  Note That 
This Film Sequence Immediate Follows the Clockwise Test. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure C.4.  Frame Extracted Near Maximum Restoring Counterclockwise Twist from 
Test Video.  Note the Buckled Graphite Rods.  
 
80 
 
 
APPENDIX D 
PHOTOGRAPHS OF POSTTEST  
OF FINAL SMART TWISTING TRUSS 
 
 
0 degrees
75 degrees
15 degrees
45 degrees
90 degrees
90 degrees
90 degrees 
45 degrees
75 degrees
90 degrees
90 degrees
30 degrees
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure D.1.  Final Iteration of Twisting Smart Posttest with Measurements of Twist 
 
Note that the nodes were color coded such that at high magnification, one can 
discern the rotation.  Also, several wires fractured in the test, so the degree of twist 
was likely more than in the posttest photograph. 
81 
 
82 
 
 
 
 
 
 
 
 
 
Figure D.2  Close-up Photograph of Twisting SMART with Color Coding Evident 
 
 
 
 
 
 
 
 
 
 
 
 
Figure D.3  Researcher Lori Prothero with the Posttest SMART 
 
83 
 
 
APPENDIX E 
REDUCED GRAVITY TESTING OF SMART IN BENDING 
 
84 
SMART Reduced Gravity Testing Proposal Abstract  
Reprinted from Microgravity University 
http://microgravityuniversity.jsc.nasa.gov/theArchives 
 
 
Auburn University  
2006  
Shape Memory Alloy Robotic Truss  
Shape Memory Alloys (SMA) belong to a special class of metals with a unique 
crystalline structure that can be heat trained to remember a specific size or shape, 
which it will return to when a heat stimulus is applied. Because of the ability to stretch 
and shrink repeatedly with strain recovery, SMA are the ideal materials for mechanical 
devices that perform repetitive tasks, or actuators. An actuator, as the name implies, is 
a mechanical device which causes something else to move.4 Actuators can be 
composed of a large number of moving parts such as gears and hydraulic pistons. The 
light-weight and simplicity of shape memory alloys have made them a great 
alternative to conventional actuator systems. SMA have the capacity to reduce the 
complexity of an actuator by replacing many moving parts with a few strands of SMA 
wire. This reduces spacecraft actuator mass, which can be significant to reducing 
launch costs. Replacing hydraulic pistons and gears with a single wire also reduces the 
vibrational settling time, or shaking, as the actuator moves, causing hydraulically 
mechanized machines to be imprecise in their movements.4 As a result, vibration is a 
major limiting factor in robotic design, especially in a space environment because it 
has fewer media to help dampen the undesired movement. Two proof-of-concept 
designs previously completed by team members show the feasibility of SMA 
actuation, both linearly and torsionally. The overall goal of Team SMART is to build 
and test two fully functional SMA actuated Remote Manipulator Systems (RMS?s) 
designed to function in a reduced gravity environment. The team will then compare 
mechanical characteristics including fatigue life, nodal deflections, total deflections, 
force, stress, and coiling phenomena between gravitational field and reduced gravity 
environments. The testing will increase the technological readiness level for SMA 
spaceflight actuated devices and will help to develop a correlation factor for linear 
actuation and coiling phenomena.  
 
 
 
 
 
 
 
 
 
 
 
 
Figure E.1.  Auburn University ?Vomit Comet? Team at Team Readiness Review 
Team Members (Left to Right):  Meghan Brown, Chris Worley, Unknown Girl Not on 
Team, Vanessa Smith, Andrew Wright, Ryan Lureck, Michael Brennison 
 
85 
 
 
 
 
 
 
 
 
Figure E.2.  Team Readiness Review Presentation with SMART Test Module at Right 
 
 
 
 
 
 
 
 
 
 
 
Figure E.3.  Demonstration of SMART Operations Inside Plexiglas Enclosure with 
NASA C-9 Aircraft in Background 
 
 
 
Figure E.4.  Unpacking and Presentation of Twisting SMART at  
Team Readiness Review 
 
86 
 
87 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure E.5.  Team Preflight Physiological Testing in Altitude Simulator 
 
 
 
 
 
Figure E.6.  Andrew Wright during Physiological Test 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure E.7.  Flight Crew and Flyers from March 9, 2006 at Johnson Space Center in 
Houston, Texas flight aboard the NASA C-9 Reduced Gravity Aircraft 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure E.8.  Ryan SMART Flyers Leureck and Vanessa Smith  
from March 9, 2006 Flight 
88 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure E.9.  Andrew Wright, Backup Flyer for SMART Received Opportunity to Fly 
with Experiment from Another University on March 9, 2006 
 
 
 
89 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure E.10.  Ryan Leureck During Reduced Gravity Flight on March 9, 2006 
 
 
 
Figure E.11.  Andrew Wright During Reduced Gravity Flight on March 9, 2006 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure E.12.  Ryan Leureck and Vanessa Smith Operating SMART During Reduced 
Gravity Parabola on March 9, 2006 
90 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure E.13.  Vanessa Smith and Ryan Leureck Observing SMART During Reduced 
Gravity Parabola on March 9, 2006 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure E.14.  Vanessa Smith and Ryan Leureck Observing SMART Through Top of 
Plexiglass Enclosure on March 9, 2006 
91 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure E.15.  Flight Crew and Flyers from March 10, 2006 at Johnson Space Center in 
Houston, Texas flight aboard the NASA C-9 Reduced Gravity Aircraft 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure E.16.  Smart Flyers Michael Brennison and Meghan Brown with SMART Test 
Module on March 10, 2006 Flight 
92 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure E.17.  Smart Flyers Michael Brennison and Meghan Brown During Reduced 
Gravity Flight on March 10, 2006 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure E.18.  Meghan Brown During Reduced Gravity on March 10, 2006 
93 
 
94 
 
 
 
 
 
 
 
 
 
 
 
 
   
 
 
 
 
 
 
 
Figure E.19.  Michael Brennison Checking Computer In Flight on March 10, 2006 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure E.20.  Meghan Brown Checking SMART Bending Truss Experiment Through 
Plexiglass Top of Experiment Enclosure on March 10, 2006.