Development and Testing of a Bi-Stable Actuator based on a High Transition
Temperature Shape Memory Alloy
Gabriel Paciotti*, Jacques-Eric Bidaux**, Hervé Girard**, Fabrice Rottmeier* and Luc Blecha*
Pyrovalves, i.e., valves actuated by a pyrotechnical device, are very often used in space propulsion systems
for their fast response, high reliability, low mass and low cost. However, they are hazardous and single use
items. Their operation induces shocks and can generate contaminants. If a slower actuation is required or
possible, shape memory-based actuators avoid the generation of shocks and contaminants while being
fully resettable and non-hazardous devices.
The goal of this project was to develop a bi-stable Shape Memory Alloy (SMA) actuator for a specific Slow
Acting Latch Valve with long life capability. The benefits of this actuator makes it interesting not only for
valve actuation but also for any device that has to be actuated in a smooth and reliable way.
This paper presents the development of a bi-stable actuator based on high temperature shape memory
Shape Memory Alloys undergo a phase transformation upon cooling from high temperature austenitic
phase to a low temperature martensitic phase. This phase transformation is the basis for the unique
properties of these alloys. At room temperature, the SMA can be readily deformed (low yield strength).
When heated above the reverse transformation temperature Af, the alloy reverts to austenite and recovers
its previous shape. This property is known as the shape memory effect. With proper selection of SMA
materials and production processes, the transition temperatures can be tailored to the engineering needs.
Figure 1. Shape memory effect principle
The team designed, manufactured and tested an actuator based on a high-temperature shape memory
alloy. The SMA actuator main features are:
• High reliability
• Light and compact
• High temperature SMA alloy (above 120°C)
• Withstands launch loads as high as 28.3 gRMS
• Made of space-compliant materials
* Almatech, EPFL, Lausanne, Switzerland
** University of Applied Sciences and Arts Western Switzerland, Sion, Switzerland
Proceedings of the 43nd Aerospace Mechanisms Symposium, NASA Ames Research Center, May 4-6, 2016
The SMA actuator developed by Almatech is based on the following operating principle: electric current is
used to heat up multiple SMA wires (in a parallel configuration) changing their crystalline structure in such
a way as to shorten them and actuate the integrated latching/de-latching mechanism. Thanks to this bi-
stable custom-designed latching device, no power is needed to maintain the actuator in one or the other
The overall dimensions of the developed actuator are shown in Figure 2.
Figure 2. Overall dimensions of the Actuator
Table 1 presents the main requirements that drove the design of the actuator.
Table 1. Main design requirements
Overall max dimensions :
100 x 60 x 60 mm
Lifetime (with ECSS factors) :
Power consumption (DC) :
<50 W (28V)
< 1.7 mm
> 2.2 mm
Actuation time (slow actuation):
5s < t <60s
Max non-operating temperature (Qual):
-10°C / 120°C
Max operating temperature (Qual):
-5°C / 100°C
Launch vibration load :
bi-stable (stable in both open and
with no power supply)
Nickel-Titanium (NiTi) alloys are the most commonly used materials for their shape memory effect. They
are off-the-shelf products and relatively low cost due to their extensive use in the biomedical and general
industrial fields. While their maximum transition temperatures (Ms, Mf, As and Af) are well adapted to these
applications, they are all under 100°C, which limits considerably their use in space mechanisms design. In
order to widen the application range of SMAs in space mechanisms, Almatech has initiated, in collaboration
with the University of Applied Sciences of Sion, Switzerland, a new development based on another shape
memory alloy, CuAlX. CuAl-based material can feature, with the adequate composition and treatments,
transition temperatures higher than 120°C. In addition, shape-memory strains up to 8-10% can be observed
with these SMA alloys representing a significant advantage over standard NiTi alloys limited to shape-
memory strains up to 6-7%. On the other hand NiTi alloys are electrically resistive and can therefore be
directly heated by Joule effect using through current as energizing source. Due to their high copper content,
CuAl-based alloys have the main drawback of being electrically conductive such that CuAl-based actuators
shall be equipped with a dedicated heating system that also needs to withstand deformations as large as
10% while keeping its full heating functionality. Higher transition temperatures and larger deformations
make in general the CuAlX more suitable for space applications, however at the cost of a more complex
SMA heating design.
Mechanically, the challenge was to establish a bi-stable design (no electrical current needed to maintain
both end positions) that is highly reliable and able to sustain the launch loads without state change. Friction,
wear, fretting and cold welding are issues that have been considered and worked around during the actuator
development phase, but the true challenge turned out to be the mechanical stability of the bi-stable
mechanism during launch, which required a particular effort.
SMA Wire Actuator Main Features
SMA wires work in traction in their straight configuration to provide their full capability while being
accommodated in parallel in the actuator to allow redundancy.
The main challenge in the heating system design was to sustain the large actuation deformation (8-10%)
produced by the CuAlXx alloy. Indeed to take full advantage of the capabilities of the selected SMA material,
the wire was equipped with the following elements:
• Wire Ends, providing the mechanical interface
• An electrical crimp for electrical back loop through the SMA
• A heating system to provide the necessary thermal power to the SMA
Figure 3. Equipped SMA Wire
A performance test highlighted the capacity of the naked wire (wire without heating system) to perform
several thousands of cycles, making the required 1000 cycles life test easily achievable for the SMA itself.
The dimensional change leading to actuation is intrinsic to the microstructure of the material and therefore
heavily dependent on the production processes. High reliability of the SMA wire can therefore be insured
by a thorough control of the production processes and characterization of production batches. The heating
system was identified as the main design challenge very early in the project and a particular effort was
dedicated to improve its design and reliability. Therefore, each wire features a redundant heating system.
In addition to this heating system redundancy, the device was designed to be fully operative and delivering
the full nominal force with a single disabled SMA wire, providing a double redundancy in the system.
The SMA wire heating system is based on the use of a geometrically adaptive external heating element
running along the full length of the actuating wires. The SMA wires being conductive, they are used as
return lines to allow the electrical powering from a single side. This avoids the use of movable electrical
connections following the extended and retracted states of the actuating wires on the opposite side.
Mechanical interface Mechanical interface (fixed end)
On the moving end, an electrical crimp makes the electrical connection (see Figure 4) between the heating
system lead and the SMA wire, while the opposite end (the fixed end) is directly connected to the electric
power supply. The heating system lead located on the fixed end is connected to the other pole of the
Figure 4. Electrical interface: (A) Flying leads after heating system installation
(B) Connection from the fixed side
The thermal interfaces of the SMA wire system plays an important role in the performance of the full system.
Although radiation is not negligible for space conditions, preliminary tests first focused on the conductive
interface design. The initial SMA performance tests showed the strong impact of the conductive interfaces
on the overall wire thermal behavior. It was observed that the conductive thermal flux significantly lowers
the end of the wire temperature (see Figure 5). This gradient induced a significant loss of the net stroke of
the sample by limiting the extent of the shape memory effect.
Figure 5. Temperature losses in the mechanical fixation
To enhance the thermal uniformity, the wire ends were thermally insulated from their supporting elements
in a second setup. This led to the design shown in Figure 6 applied at each end of the SMA wire. While on
one hand thermal insulation is desired to maximize the performance of the SMA wire by a uniform heating
over its full length, heat rejection is required after SMA wire heating such that it can come back to its initial
state and be actuated again with a new thermal cycle. Therefore, a compromise has to be found on the
thermal insulation design that considers the trade-off between SMA wire heating efficiency and required
actuation cycling times.
Figure 6. Thermal insulation at mechanical interface
Figure 7. Wire End crimped to the SMA wire sample
The lifetime performance of a SMA being limited by its exposure to over stroke, overstress and overheating
conditions, the powering of the SMA wires needed to be controlled to meet the required lifetime
performance of 1000 cycles. In order to do so, an electrical switch has been integrated in the design to cut-
off the powering of the SMA wires as soon as the system is latched and a mechanical end-stop has been
integrated to guaranty a repeatable rest position after cooling down of all SMA wires.
The switch is integrated on the backside of the actuator. A switch lever is pushed down by a Spring Blade
as illustrated in Figure 8. The switch is released once the Pusher rod is in latched position.
Figure 8. Switch actuating system
When the SMA Wires are activated, the Pusher rod travels forward. Once the adjustable nut supported by
the threaded rod touches the Spring Blade, it shuts off the electrical supply and heating of the SMA wires.
Adjustable Nut (self-locking)
Figure 9. Fully assembled SMA Actuator
SMA Actuator Test Sequence
The test campaign consisted of the following test sequence:
a. Cycling tests at SMA wire assembly level
b. Cycling tests at mechanism level
c. Mechanism vibration test
The first test aims to verify the correct actuation of a wire alone before integration into the actuator. It allows
to ensure its capability to actuate under the required load and over the required number of cycles. Full
performance test results of the developed Engineering Model were prioritized over the verification of its
structural resistance to launch loads, such that these tests were performed at the end of the campaign and
followed by final functional tests. The corresponding test results are presented hereafter.
SMA Wire Cycling Test Results
The goal of the life test was to demonstrate the maximum number of actuation cycles that can be performed
by an equipped SMA wire. The pass/fail criteria considered was the successful achievement of more than
1000 actuation cycles with 2.7-mm stroke under 70-MPa load.
The schematic drawing of the test jig is shown in Figure 10. The equipped wire samples were mounted to
the fixtures of the test jig and a weight was attached to the lower fixture. The test setup is such that the
initial position of the weight can be either in contact with the platform (load relief) or hanging on the SMA
wire (free load). The SMA wire is actuated using the developed heating system using a constant current.
The wire shrinks when actuated and lifts the weight in a loaded configuration. The displacement is
measured using an LVDT transducer. As soon as the required displacement is achieved, the current is
switched off. The weight then gradually returns to its initial position as the wire cools down and a new cycle
is initiated. The test finishes when the predefined number of cycles are performed. The SMA wire
temperature is monitored continuously using a type K thermocouple.
Figure 10. Schematic drawing of the equipped SMA wire test jig
The tests were performed starting from two initial positions: free load (weight hanging) and offloaded (weight
in contact with the platform). The position of the upper fixture was adjusted using a micrometric screw.
Two tests were performed with the following load conditions:
• 40 MPa (30.62 N), to verify the ability of the SMA to reset (recover the original position)
• 70 MPa (51.12 N), the maximum determined stress for infinite cycling
Figure 11. Cycling under the load of 70 MPa
Figure 11 shows a uniform cycling demonstrating the adequate behavior of the wire at 70 MPa. The same
test was performed at 40 MPa demonstrating its ability to reset fully at this stress level.
Figure 12. Cycles 7646-7657, load 70 MPa, stroke 2.7 mm, current 1.94 A
By the end of the test, 7657 cycles were successfully performed as illustrated Figure 12. During the cycle
7658, a fatigue failure occurred close to the middle of the wire as shown in Figure 13.
Figure 13. Equipped wire after wire failure
Actuator Cycling Test Results
In order to study the cycling and time-dependent properties of the actuator, the following test was
• Loading force = 96 N (reduced from the theoretical 175 N to ensure SMA cycling survival and
verify the heating system over the full lifetime requirement)
• Power supply : 1.79 A, 27.8 VDC, 49.8 watts
• Cooling time between cycles = 15 seconds (to ensure that the actuation time at the next cycle will
not be affected by an incomplete cooling at the previous cycle)
• Heating time limit = 15 seconds (overheating risk)
• Standard ambient temperature and pressure conditions
The actuator survived the required 1000 cycles and switched its state (latched or released) automatically
at each cycle. The cycling test of the actuator was fully successful as illustrated in Figures 14 and 15
showing the four initial cycles and the last four cycles.
Figure 14. Cycles 1-4
Figure 15. Cycles 1054-1057
Actuator Vibration Test Result
Resonance search and random vibration tests were performed in order to ensure adequacy of the actuator
design to the required launch loads. The tests were performed in axial and radial directions.
Figure 16. Axial vibration of the Actuator
Figure 17. Transverse vibration of the Actuator
The overall acceleration of the applied random vibration profile was 28.3 gRMS and the total test duration
was 3 minutes. The vibration spectra acquired during the tests are shown in Figure 18 and Figure 19.
Figure 18. Random vibration spectrum in the axial direction
Figure 19. Random vibration spectrum in the radial direction
The SMA actuator successfully sustained both axial and transverse random vibrational loads with overall
acceleration of 28.3 g RMS and duration of 3 minutes each. No changes in the resonance search results
were observed before and after the random loads application showing that no structural failures have
occurred. A post-test functional verification showed that the actuator remained fully operational after the
test. It proves that the design is able to survive the required launch vibration loads.
Synthesis of Actuator Characteristics
Table 2. Actuator main characteristics
Overall dimensions :
89 x 44 x 44 mm3
100 x 60 x 60 mm3
(stopped, no damage)
> 1000 cycles
SMA wire lifetime:
> 7657 cycles
49.8 W (27.8V, 1.79A)
< 50W (28V)
Verified output force:
6.5± 0.4 s
5s < t < 60s
5s < t < 60s
Max operating temperature:
Launch vibration load:
*The quality of the produced CuAlX monocrystalline wire samples was not sufficient to reach the cycling
stability under the required force. It was therefore decided to reduce the test load at 96 N (from 165 N) for
the test campaign. Recently, industrial quality level CuAlX wires are commercially available and should
meet the design load of 175 N.
A bi-stable actuator based on a high transition temperatures Shape Memory Alloy has been successfully
designed, manufactured and tested. This mechanism, able to provide slow-acting pushing and pulling
forces, showed great repeatability over a high number of actuating cycles demonstrating high overall
reliability of the selected design.
One major challenge was to design a SMA heating system compatible with the use of high transition
temperature SMAs that are highly conductive and have shape-memory strains up to 10%. A novel heating
system has been developed and successfully tested on single wire setups as well as on a fully integrated
Through this project, the technology based on high transition temperatures SMA reached the readiness
level TRL-4. Further material characterization tests on industrial quality CuAlX wire as well as additional
environmental tests are to be performed to reach TRL-5.