Shape memory alloy gripping device

ABSTRACT

This project presents the design and development of a shape memory alloy gripping device. An SMA wire actuated gripper was designed and test fabricated with conventional milling machinery. The design took advantage of the small linear displacement of the SMA wire to convert it into angular movement of the gripping jaws.

As part of a micro-assembly system, micro-grippers are very important as they cater for all the actual work being undertaken. Each of these grippers consists of an actuator providing the actuation for the gripping action. SMA wire actuated grippers, has the advantage of smaller size and the abilities of exerting ideal gripping forces for micro actuation applications.

In this project, the design of the gripper focuses on simplicity as well as performance. The SMA actuation is provided by pulsing electric current from a driving circuit. Through this method, the risk of the SMA being overheated can be reduced and yet providing sufficient power for useful gripping task. This would ensure the gripper with a longer and lasting actuation.

Experiments conducted on the working prototype are aimed to gauge the gripper’s performance in terms of its working response as well as reliability. These include the generated force, level of excitation, reaction time, gripper’s characterization as well as the cyclic performance. The outcome from these tests shows satisfying results. The gripping force achieved was in the range of 70 mN to 500 mN, ideal for micro-applications. This force can be varied by tweaking the driving current through a variable resistor. In terms of reliability, millions of cycles were accomplished without any deterioration in performance. This proves that the concept of the design is feasible and the gripper is able to support continuous long operation of work before any replacement of parts is required. One drawback of this SMA actuation is its relatively slow actuation speed. But in a region of under 1 second, the gripping speed can be considered satisfactory given that the application involved does not require lighting fast reactions.

ACKNOWLEDGEMENTS

The author would like to express his sincere appreciation to Associate Professor Zhong Zhaowei, the supervisor of this Final Year Project for his valuable advice and guidance throughout the duration of the project.

The author would also like to extend his gratitude to technical staff members of the Mechatronics and Control Laboratory especially Miss Lee Koon Fong as well as Mr. Teo Ah Khoon of the Mechanical Service Workshop for their assistance and guidance in various ways towards the successful implementation of this project.

Last but not least, special thanks are extended to the author’s family and friends, for their care and support throughout the duration of the project.

Chapter One

INTRODUCTION

Today, micro-assemblies in product and devices are nothing new. This can be seen in the smaller and more compact products available in the market. This forces an increasingly number of high precision components to be crammed into ever shrinking packages. Technologies in the micro-scale have been developed over the years for consumer as well as specialized applications such as in the field of electronics, optics, medicine and biology covering areas such as diagnostics, drug delivery, tissue engineering and minimally invasive surgery (MIS) [1]. While considerable development have been made in the fabrication of micro-parts, unfortunately, the assembly of these micro-systems still account for a substantial portion of the final cost.

In recent years, researches have been carried out on micro-assembly, in particular micro-gripper or micro-manipulator which involve a wide range subjects including their fabrication procedure, the structural design, actuation, kinematics and control. In this project, the author would study the use of shape memory alloy (SMA), muscle wire (MW) in particular as a replacement of the more conventional piezoelectric motors as well as hydraulic pistons in the design and fabrication of a gripping device for handling components in the miniature-scale.

Background

Muscle wire, made from Nickel-Titanium (also known as NiTiNOL) is a revolutionary alloy in the form of wire with the special ability to recover its original shape when subjected to a certain level of heat. This phenomenon is known as the shape memory effect (SME). It arises when the wire is heated above a certain transition temperature, transforming its crystalline phase from martensite to austenite phase. This can be accomplished by simply applying a voltage drop across the wire which induces current flow and resulting in joule heating.

Apart from ease of actuation, some other advantages of these SMA wires include its incredibly small in size as well as its high force to weight ratio. Its limitations though, include its non-linear behavior and small contraction ratio (8 % maximum).

Objectives

The objectives of this project are to study the use of SMA wire as an actuator and subsequently design and fabricate an SMA wire actuated working prototype gripper that has the ability to handle devices in the range of 100 to 500 micro-meters.

Scope

The scope of this project encompasses the study of various actuation techniques, designs, fabrication and testing of an SMA wire actuated gripping device.

Chapter Two

LITRITURE REVIEW

Micro-Assembly System

Micro-assemblies refer to the assembly of components in the micro-scale which includes the handling of components in the micro range. One of the basic challenges in precision micro-assembly is the need for very high accuracy (often in the sub-micron range) over a large range of motion [2]. It shares many common aspects with traditional robotics assembly such as positioning of manipulator, velocity, jerk and force control, tactile feedback, task planning, collision avoidance, grasping, part orientation etc.

However, the design and fabrication of micro-assemblies cannot be approached by traditional methods and technologies as they open completely new problems which are addressed by a new field of engineering known as micro-engineering. Micro-engineering aims to develop technologies and methods in design, fabrication and assembly of components or products with size from some mili-meters down to a few microns.

For a typical micro-manipulation workstation for biological objects or mechanical micro-components, it has a modular structure and consists of several interchangeable and application dedicated end-effectors (micro-grippers). The workstation is made up of several subsystems which includes a visual subsystem, a micro-positioning subsystem (with micro-actuators and micro-sensors for the accurate positioning of the micro-gripper), a control unit and a human to machine interface (HMI) which consists of software simulation tools to aid in the micro-assembly process. This is depicted in the block diagram as shown in Figure 2-1. The end-effector or the micro-gripper is an important component of the micro-assembly system which consists of tools to perform the task of manipulating the biological objects or mechanical micro-components.

Micro-Grippers

Since micro-mechanical systems have become more complex like micro-optical devices, micro-fluidic components or other hybrid micro-systems, it has become a research field of great importance. The production of micro-systems sets high requirements for the process of micro-assembly due to the small size of the micro-components to be mounted in hybrid micro-systems (such as micro-lenses, optical fiber, micro-tubes) and the low tolerances involved. This leads to very high requirements of accuracy within the assembly process. These accuracies cannot be achieved by conventional assembly equipment. Therefore the demand of micro assembling tools such as high precision robots and micro-grippers have increased dramatically. As the name implies, a micro-gripper is an end-effector designed to handle very small components for the task of assembly or pick and place. In the medical or biology field, specialized micro-grippers are designed to manipulate cells or bio-batteries for implantable devices.

A typical micro-gripper for the lower end of the scale components is made of thin plate silicon or stainless steel, designed as a so-called compliant mechanism such that the conventional bearings have been replaced by flexure hinges. Flexure hinges can be described as regions with decreased stiffness, generally made by reduce in the cross section area as shown in Figure 2-2. Actuation is typically by piezoelectric motor, SMA foil (laser-cut and mounted on the gripper) [4, 5] or SMA surface deposition [6]. The gripper is generally made by lithography-based techniques, a technology widely used for integrated circuits fabrication. It can also be made by precision micro-machining, EDM or laser cutting methods.

For higher-end-scale components (500 microns to 1 mm), forceps type of design is typically used as shown in Figure 2-3. In this case, the angle of opening of the gripper is normally bigger. Actuation can be achieved by motors, pistons, solenoids, cable actuation or SMA in various forms (bar, spring, wire) [7, 8].

Actuators and Gripper Design

In the design of a micro-gripper, there are several ways that the actuation can be achieved. This can be done by miniature electrical motors, piezoelectric motors, hydraulic pistons, miniature solenoids [9], cable actuation or SMA. In SMA, the actuator can be in several forms, namely bar, spring, piston or wire.

In a typical gripper design with motorized actuators, a complete gearing system is a must. Depends on the degree of freedom (on the gripper’s jaw) required, one or two miniature motors are normally used. Two motors are used if independent jaw control is desired. The gearing systems normally include a set of planetary gears coupled to a set of bevel gears to transmit the angular motion to the jaw. This kind of design presents some problem due to the reversible motion of the jaw. It results in loosening of grasp when force is applied to the jaw. Furthermore, problems associated with the motorized actuator include vibration, additional inertia as well as backlash in the gearing couplings [10].

In a gripper with hydraulic piston actuation, the piston drives a shaft which is linked to the set of jaws. When activated, the piston acts against a counter force spring to open the gripper. The main drawback of this design is the requirement of a fluid feeding system which complicates the arm architecture [7]. This disadvantage of the design can be overcome with an SMA type piston known as Electric Piston [11]. However, Electric Piston presents yet another set of problems due to its rather bulky size and relatively slow response.

Using Shape Memory spring or Ni-Ti spring in particular is very popular method of linear actuation. Here, the Ni-Ti spring acts opposite a counter force steel spring separated by a disk connected to a shaft. The shaft is linked to a set of jaws enabling the gripper to open when the Ni-Ti spring is activated. The drawback of this design is the long cooling time required to deactivate the Ni-Ti spring in order to close the gripper [7].

Another form of SMA that can be used in a gripper is the SMA wire. Due to the short contraction length of about 3 to 5 %, the design would have to incorporate special mechanism such as pulleys to amplify the motion. A spring would have to be put in the jaw to enable it to close once the SMA wire is deactivated [7, 12].

Shape Memory Alloy

The first SMA was discovered in 1932 by a Swedish researcher name Arne Ölander. The shape recovery abilities of a gold-cadmium (Au-Cd) alloy was discovered and noted for its potential in creating motion. In 1950, L.C. Chang and T.A. Read from Columbia University (New York) showed that the SME could be used to perform physical work. This finding generates great interest in the research community that led to the discovery of indium-titanium (In-Ti) alloy with similar shape recovery capabilities. However, the costly gold and indium as well as the toxic nature of cadmium in these early SMAs have hindered further research effort [13].

In 1963, a team led by W.J. Buehler at the U.S. Naval Ordnance Laboratory (NOL) discovered the SME in a nickel-titanium (Ni-Ti) alloy, thus the name NiTiNOL. This Ni-Ti alloy was non-toxic, less expensive and had a better deformation-to-recovery ratio. This finding sparked a new wave of interest that led to the discoveries of other shape changing alloy such as copper-aluminum-nickel (Cu-Al-Ni), copper-tin (Cu-Sn), copper-zinc (Cu-Zn), copper-gold-zinc (Cu-Au-Zn), copper-zinc-aluminum (Cu-Zn-Al), iron-platinum (Fe-Pt), nickel-aluminum (Ni-Al) and manganese-copper (Mn-Cu). Among these SMAs, Ni-Ti and Cu-Zn-Al turned out to be the best due to their strength, low cost and large changing capabilities [13].

Principles of SME

The formulation and processing of SMA plays a big role in the behavior of the final product. For nickel-titanium alloy, it typically contains nearly equal atomic amount of nickel and titanium. Varying this equation by less than one percent could lead to a very significant change of its transition temperature.

The SME in SMA is caused by a solid state phase transformation. Above a certain transition temperature, the SMA is austenitic as shown in Figure 2-4(a). Below that temperature, it is martensitic (Figure 2-4(b)). In austenite phase, the alloy is hard while it becomes relatively soft and behaves elastically in the martensite phase.

If no force is applied during heating and cooling, the microstructure of the SMA would change without a macroscopic noticeable shape change. Here, as illustrated in Figure 2-4(b), the grains of the material form a twinned structure. This can be examined in the martensite’s stress-strain curve (as shown in region 0-1 of Figure 2-5) where the material appears in the elastic region.

In the martensitic condition, the alloy can be easily deformed into a new shape by non-conventional atomistic mechanism. Here the crystalline microstructure de-twin as the grains re-orientate (Figure 2-4(c) and Figure 2-5, region 1-2). The device remains in the new shape, as long as the temperature is kept constant. At this stage, the second elastic region occurs from region 2-3 of Figure 2-5 and if the alloy is subjected to further strain, permanent plastic deformation would occur. Heating above the transition temperature would cause the alloy to transform back to the austenite phase thereby restoring its original shape.

Super-elasticity is the result of a stress induced transformation from austenite directly into deformed martensite at a temperature slightly above the transition temperature. When the deforming stress is released, the alloy would spring back into its stable austenitic state and shape [14].

The change within the crystalline structure of an SMA during SME is a thermodynamically irreversible process due to internal friction and structural defects. This causes the temperature range for the martensite-to-austenite transformation, which takes place upon heating, is somewhat higher than that for the reverse transformation upon cooling. The difference between the transition temperatures at heating and cooling is called hysteresis as illustrated in Figure 2-6.

At point 1, the alloy is 100 % martensite. During heating, the alloy follows the lower curve and starts to transform into austenite once the temperature reaches AS. At the temperature of AF, the alloy is 100 % austenite. At the cooling phase from point 2 to 1, the alloy starts transforming from austenite back to martensite at the temperature of MS. At the temperature of MF, the alloy is back to its original structure of 100 % martensite. This temperature hysteresis can be translated directly into hysteresis in the strain-temperature curves as shown in Figure 2-7. The behavior of hysteresis in SMA has made the development of SMA actuators increasingly challenging. Some researchers have chosen to simplify the problem by creating a dynamic model where the phase transition temperature is the same for heating and cooling, ignoring the effect of hysteresis [14].

The complexity in SMA models present great difficulties to mathematically design controllers for SMA actuators. These difficulties are mainly due to the three following factors: [14]

  • nonlinear dynamics of SMA
  • difficult to identify the state of SMA
  • SMA actuation is sensitive to the change of environmental temperature

Advantages and Drawbacks in SMA Actuation

SMA offers important advantages in actuation mechanisms. This is summarized as followed.

Simplicity, compact, and safety of the mechanism:

SMA actuator can be as compact as a single piece of wire activated by electrical current. Its stroke and force can be easily modified by the selection of a suitable form such as SMA wire or a SMA spring. Additional parts such as reduction gears are not required. Hence, the use of SMAs can result in a simplified, more compact and more reliable final product.

Clean, silent, spark-free and zero gravity working conditions:

Since SMA activation is frictionless, the problem of dust particles due to friction can be avoided. Conversely, a dusty environment has no influence on the action of SMA elements. In the absence of additional vibrating parts, the activation of SMAs is almost noiseless. While no high-voltage or electrical switches are required, SMA actuators can work completely spark-free allowing them to operate in highly inflammable environments. SMA actuators can be controlled in such a way that accelerations of the order of only a few mG are generated making it extremely suitable for space applications where even small accelerations can influence the global movement of satellites [15].

High power to weight ratio:

The SMAs are extremely attractive in micro-actuator applications due to its powerful strength per unit weight [15]. In addition, its strength can also be amplified by connecting numerous SMAs in a parallel configuration. This is an added advantage especially when space and weight are major constraints in the design of the product.

However, there is the present of some drawbacks that must be considered when implementing SMA actuators.

Low energy efficiency:

The conversion of heat into mechanical work produces a very inefficient result (less than 10 % efficiency) [15]. To overcome this problem, complex control strategies is required

Single way force:

SMA elements can only provide force and displacement in one direction. An SMA wire that compresses when heated does not expand without the aid of an external force when it cools down. To overcome this problem, a bias mechanism can be employed. However, the use of bias mechanism would require space and increase the weight of the device, resulting in a more complex design. Furthermore, the net output force of the device would also decrease due to the opposing force from the bias mechanism.

Heating and cooling restrictions:

SMA actuators can be heated in several ways, radiation or conduction (as in thermal actuators) and by inductive or resistive heating (as in electrical actuators). For a fast and homogeneous response, resistive heating offers the most attractive solution and is therefore widely used [15]. However, the response speed is mainly limited by the cooling capacities due to the environment factor.

Degradation and fatigue:

The reliability of shape memory devices depends on its global lifetime or cyclic performance [15]. Time, temperature, stress, strain and the amount of cycles are important external parameters. Internal parameters that can have a strong influence on the lifetime are the alloy system, the alloy composition, the heat treatment, and the processing of the SMA.

Complex control:

SMAs demonstrate a complex three dimensional thermo-mechanical behavior with hysteresis [16]. Moreover, this behavior is influenced by a large number of parameters. There are in general no direct and simple relations between the temperature and the position or force. Therefore, accurate position or force control by SMA actuators requires the use of powerful controllers and the experimental determination of complex data.

Applications of SMA

SMAs can now be found in a wide range of applications and products that take advantage of their thermal recovery and super-elastic properties. Among these applications are in the field of medical and engineering.

In the medical field, Ni-Ti alloy received much attention due to its high compatibility with living tissue and low risk of rejection when implanted. Since the 1970’s, researchers have been toying around with ideas of using Ni-Ti alloy to activate artificial organs. In fact, in 1971, P.N. Sawyer and M. Page of New York built an experimental artificial heart with electrically triggered Ni-Ti wire as actuators. However, it failed to go beyond that stage due to relatively slow contraction-expansion rate of Ni-Ti alloy. This however did not stop Dr. Morris Simon from developing the expandable blood clot filters. Tiny web of Ni-Ti alloy in collapsed form is inserted into a patient’s vein or artery. When in place, the web is expanded by the heat of the body trapping potentially stroke-causing blood clots. Some other successful implementation of Ni-Ti alloy can be found in surgical tools as well as in dentistry where SMAs is used to develop braces and tooth alignment structures [13].

In the field of engineering, it covers aerospace, military, automotive as well as industrial level of applications. The most successful implementation to date is the use of SMA in pipe couplings. Installed at low temperature, the couplings shrink into place and provide highly reliable, tight and flawless connections for aircraft hydraulic lines. Some other applications of SMAs in this field include heat engines, electrical connectors, thermostats, sensors, valves as well as robotics to a certain extend [13].

SMA can also be found in consumer products such as spectacle frames, golf clubs, antennas, toys and etc. More recently, SMA has been used in the development of Micro-Electro-Mechanical Systems (MEMS) devices. Components such as micro-valves in micro-fluidics [17], switches, mirrors, micro-positioners, micro-robotic systems are being made with SMA for various industry applications.

Chapter Three

MECHANICAL DESIGN

Given the limitations of SMAs, numerous considerations had been taken in the design of the gripper. Among them, the design must make use of the little displacement of SMA wire to translate into the gripping action while keeping the SMA contraction under the recommended limits. In addition, the gripping force achieved should be sufficient to perform meaningful task without damaging the parts to be handled. Furthermore, the gripper should be able to execute its tasks consistently and reliably throughout its lifetime. This chapter presents several designs of gripper at the conceptual stage as well as the final stage.

Conceptual Design

Design A

Design A as depicted in Figure 3-1 is based on the concept of a compliant design. The one piece design requires the use of precision micro-machining or laser cutting technique to fabricate. Material can be of a piece of a silicon or stainless steel plate. Two screws are used to fasten the gripper to the base while the center screw is used to fasten one end of the SMA wire (MW) to the center portion of the gripper. The other end of the SMA wire is to be tightly fastened to the base with another screw (not shown in figure). Both ends (of the SMA wire) are then connected to an electrical actuation circuit which would provide electrical current to activate (heat up) the wire.

The gripper is designed such that once the SMA wire is activated, it would contract and the center portion of the gripper would be pulled towards the base as shown in Figure 3-2.

When this happens, the jaws of the gripper would deflect inwards and thus able to grasp on parts to be manipulated. To release the gripper, one simply needs to cut off the electrical current flow into the SMA wire by performing an open circuit. The elastic behavior of the material would enable the deflected gripper to spring back to its original position.

Design B

The construction of gripper in Design B is fairly simple. A small stainless steel tweezers is form and fitted into a hollow cylinder body (holder). One end of a SMA wire (MW) is fastened to the back of the tweezers by a screw as shown in the section view in Figure 3-3.

The other end of the SMA wire is tightly fastened to the cylinder holder (not shown). Electrical terminals from an actuation circuit are connected to both ends of the wire. The tweezers is carefully designed so that when the SMA wire contract by about 5 %, it would be forced to close by the inner wall of the cylinder holder as shown in Figure 3-4. When the electrical current is cut off, the SMA wire is relaxed and the springy nature of the tweezers would force itself to open naturally.

Though simple and easy to construct, a potential problem with this design is that the amount of work put in to force the tweezers to close could be tremendous due to the large friction with the inner wall of the cylinder holder causing it to be inefficient.

Design C

The gripper of Design C as depicted in Figure 3-5 is made of angle brackets with hinges to the housing. The ‘moving-block’, together with the spring acts as a stopper to limit the opening of the gripper. The plug has a unique shape to force the gripper to close. It is connected to one end of the SMA wire with a screw.

When the SMA wire is activated, it would pull the plug towards the direction of the arrow as shown in Figure 3-6. This would compress the spring and subsequently close the gripper.

When the SMA wire is relaxed, the spring would react and force the gripper to open. Step and slots are created on the inner plate of the housing to allow the plug and moving block to sit in and slide about a predefined area and direction. A plan view of the design concept is as shown in Figure 3-7.

Final Design

Design Version 1

The elements of the final design were derived from the concepts of Conceptual Design C. The gripper at this stage was fabricated from standard aluminum bar and plates. As shown in Figure 3-8, the two jaws of the gripper were pivoted to the base with side walls and a retaining plate constraining them to an opening of approximately 1 mm.

Actuation is to be provided by a short piece of SMA wire with both ends terminated to a piece of perf board by a pair of screws and nuts. The SMA wire formed a loop around a protruding screw attached to a sliding unit. The sliding unit is designed with an opening slot at one end providing a linkage to both jaws so that when the SMA wire is activated, it would pull the jaws close simultaneously. As depicted in Figure 3-9, a pair of bias springs was mounted at the back of the jaws to aid the recovery of the gripper to its initial state when the SMA wire was deactivated. The base of the gripper was design with grooves and steps to accommodate the insertion of the springs and the stopping block.

Design Version 2

The prototype fabricated in Design Version 1 works when tested. But it faced some inconsistency problems especially during the recovery of the gripper to its initial state (opening).

First of all, the sliding unit is not properly guided along its intended path. To rectify this problem, the design of stopping block was modified with C-slots integrated at both sides (as shown in Figure 3-10) to allow the sliding unit to slide through it. A third spring was added at the back of the sliding unit (as shown in Figure 3-11) to ease its recovery motion and effectively prevents it from getting jammed in the C-slots.

Another notable change in Design Version 2 was the material used to fabricate the base of the gripper. Here, acrylic was chosen to replace aluminum as a prevention of any possible short circuit or leakage of current during the activation of the SMA wire. As shown in Figure 3-12, the area where the gripper jaws sit was also redesigned to make it into a single piece, eliminating the need for a retaining plate. This was achieved by cutting a through slot just big enough for the pair of gripper jaws to sit in. With changes done in Design Version 2, the closing-opening action of the gripper (as shown in Figure 3-13) can be achieved without any problem.

The assembly of this final design was fairly simple. As shown in Figure 3-14, the jaws of the gripper were first inserted into the slot on the base unit with the two bolts acted as pivots. The stopping block (with C-slots) was then mounted on with the two bias springs in placed connecting to the rear end of the respective jaws. Next, the sliding unit was glided through the C-slots before being attached to the jaws through two mounting bolts. The stopping block (U) was then mounted on with the third spring sitting in the U-slot and connecting to the rear end of the sliding unit. Finally, the perf board (with the driving circuitry) was installed with the SMA wire carefully attached and connected to the sliding unit. Detailed drawings of the Final Design can be obtained in the Appendix A section of this report.

Chapter Four

ELECTRONICS DESIGN

In the design of the driving circuit, the main concern was for it to support the actuation of the SMA wire without overheating it. In addition, the power management of the circuit should be efficient and able to meet the MW manufacturer’s recommended level. This chapter presents several electronics circuitry designs for the SMA wire excitation at the conceptual stage as well as the final stage.

Conceptual Design

The excitation of the SMA wire can be done in various ways. Among them, electrically driven method is the most practical solution. An ideal driver circuit is one which is able to keep the current stable and well below the SMA wire’s overheating limit. A simple SMA wire driver circuit as shown in Figure 4-1 can be assembled by connecting the SMA wire in series with a current limiting resistor to the power supply (batteries or DC supply).

With the voltage from the power supply fixed, the desired driving current can be obtained by adjusting the resistance of the resistor. This way, the SMA wire is prevented from overheating, thus prolonging its lifetime. However, this method has its drawback as it uses power inefficiently where excess power not required by the SMA wire is “burn up? in the resistor.

Design A

In this circuit as shown in Figure 4-2, the voltage regulator, U1 acts as a current limiter controlling the excitation current entering the SMA wire. When the switch SW1 is closed, current flows through resistor R2 to the base of transistor Q1, resulting to a “short circuit? between the collector and the emitter. This enables the controlled current from the voltage regulator to flow through resistor R1 and subsequently activating the SMA wire. With the correctly chosen value of R1, the circuit would protect the SMA wire from overheating. However, it is inefficient for shorter lengths of SMA wire, as it has to burn off excess power not required by the SMA wire.

Design B

The driving circuit proposed in Design B (as shown in Figure 4-3) is basically an astable oscillating circuit constructed from a 555 timer. The high and low logic of the oscillation can be controlled by setting the values of resistors R1 and R2 as well as capacitor C1. By setting a short cycle time of oscillation, with a slightly less than 50 % of duty cycle, the circuit should be able to protect the SMA wire from overheating.

When the output pin 3 of the timer is in logic low, it acts as a current sink, allowing current to flow from the DC supply and subsequently activates the SMA wire. When the logic high comes about, it stops the current flow (due to the absence of potential difference between DC supply and pin 3) thus allowing the SMA wire to cool down a bit. Before the SMA wire physically recovers, it is heated up again with a logic low at pin 3.

This driving circuit has the overheating protection capabilities. However it is not very efficient as excess power is burn up in the resistor R4.

Design C

In this driving circuit, the idea of a pulse width modulation (PWM) is utilized [18]. As shown in Figure 4-4, it is made of two oscillators constructed from 2 units of 555 timers. One of the oscillators acts as a trigger while the other provides the PWM signals. The triggering cycle is fixed with the proper selection of components R1, R2 and C1 in the first oscillator. It is desired that this triggering cycle be kept less than 1 second so that the performance of the SMA wire is more stable without any jerking effect. The second oscillator controls the duty cycle of the PWM. It can be adjusted by varying the resistance of VR1 and this effectively regulates the amount of power and heat generated in the SMA wire.

Final Design

The final design of the electronics driving circuit was derived from the combination of Design B and Design C. As shown in Figure 4-5, the output of the 555 timer was used to source the SMA wire directly. This was achieved without a current limiting resistor as the 555 has a maximum output current of about 200 mA, which is within the recommended region of excitation current given by the manufacturer (Mondo-Tronics).

When the push button switch SW1 is activated, the combination of resistors R1 and R2, and capacitor C1 provides an oscillation output averaging at 2 to 3 Volts as shown in Figure 4-6, with a duty cycle of 70.9 % at a frequency of about 2.3 Hertz. This simply means that for every cycle of oscillation, about 0.31 second was utilized to generate heat in the SMA wire. This was followed by 0.13 second of relaxation before the cycle was repeated.

With this circuit, ample time is given to allow the SMA wire to cool down a little before being heated up again. This prevents it from overheating thus prolonging its useful lifetime. The duty cycle of about 70 % coupled with the short cycle of oscillation (0.44 second) provides the gripper with a steady grip without any jerking effect.

The detailed driving circuit computation is presented as followed. The objective of this exercise is to obtain the appropriate values for resistors R1 and R2, as well as for capacitor C1 that produced a pulsing signal with at least 70 % duty cycle.

Chapter Five

DESIGN OF EXPERIMENT

This chapter presents the methods and setup of the tests conducted on the prototype gripper.

Gripping Force Test

The aim of this test is to determine the gripping force of the prototype gripper with respect to the excitation current of the SMA wire. For the purpose of this test, the original driving circuit presented in Chapter 4 was modified with the addition of a variable resistor (trimmer), VR1 connected between output pin 3 of the 555 timer and the SMA wire as shown in Figure 5-1.

The measurement was conducted using an innovative thin film force sensor known as FlexiForce [19] as shown in Figure 5-2. This device has a linear conductance behavior with respect to the applied forces. As conductance in Siemens is the reciprocal of resistance in ohms, the resistance of FlexiForce was taken as the unit of measurement. An important procedure before the test was to calibrate the readings of the FlexiForce with known forces. This was achieved by applying known masses of a few grams up to about 100 grams on the force pad. The resistance readings of each mass were taken and a linear line chart of force versus conductance was generated in MS Excel.

The FlexiForce sensor was calibrated with known masses in the range of 3 to 130 grams. The corresponding resistances were taken as shown in Table 5-1. Subsequently, the collected data was utilized to produce charts as shown in Figure 5-3 and 5-4 to test for the force-conductance linearity behavior of the sensor.

With the availability of the linear force versus conductance chart as shown in Figure 5-4, a resistance-force relationship was established. This relationship was then used to compute the gripping force of the gripper, given the resistance which was probed from the terminals of FlexiForce.

With the calibrated FlexiForce, the gripping force of the prototype gripper can be measured. For all available sizes of the SMA wire (37 mm, 50 mm, 100 mm, 150 mm and 250 mm), the respective desired excitation current was adjusted (with the aid of an amp-meter) by tuning the trimmer. For every indicated current level, the gripping force was measured. This was accomplished by probing the resistance of FlexiForce while allowing the gripper’s jaws to grip on the center of the force pad. The desired gripping force was then computed from the recorded resistance based on the relationship obtained during calibration. The result of the test was recorded and discussed in Chapter 6.

Gripper Cyclic Performance Test

The test discussed in this section was designed to examine the cyclic life of the SMA wire in the prototype gripper. The idea was to put the gripper through a series of repeated open-close cycles and approximate the number of successful cycle before the permanent failure of the SMA wire occurred.

The test was conducted using a modified version of the original driving circuit as shown in Figure 5-5. Here, an addition astable oscillation circuit was used to trigger the original circuit (as depicted in Figure 5-6) so that it performs the test in its original designed condition. Resistors R1 and R2 and capacitor C1 was chosen so that the gripper would close for about 1 second and open for another 1 second.

With the period of a cycle known, the number of cycle can be approximated by monitoring the total test time before failure. This tedious calculation was made simpler with the aid of MS Excel spread sheet in recoding and computation of the total cycle time. Here, the parameters required for the computation include the start date, start time, end date, end time and the period. In the MS Excel spreadsheet, the start date and time was recorded each time a continuous test was conducted. When the test was stopped for inspection, the corresponding end date and time was also registered. With the start and end information available, the time in seconds for each continuous test was computed. The number of cycle achieved in each test was then computed by division over the period. This simple computation can be summarized into the following equation.

Number of Cycles,

Where tdifference, is the difference in seconds taking into the account the number of days which has elapsed. While the constant period of 2 seconds, was obtained from the oscillation period of the driving circuit. This simply means that it takes 2 seconds for the gripper to perform one cycle of close-open action.

The algorithm used to compute the time difference in MS Excel format can be represented as followed. In the equation, the first term produced the difference in time without the day difference consideration, while the second term produced the difference in days which has been converted into the (hour:minute) format.

With a compatible format of time difference available, the number of cycles, ncycle can then be computed by a division over the period as shown in the first equation presented in the previous page.

Gripper Characterization

The objective of this section was to characterize the performance of the prototype gripper with respect to its input. Here, two attributes of the gripper design was measured, they were the gripper tip displacement and the power consumption of the driving circuit.

Tip Displacement Measurement

The aim of this test was to determine the displacement of the gripper’s jaws with respect to the excitation current from the driving circuit. This was achieved by measuring opening of the gripper jaws with increase in the SMA excitation current. For each current level, the opening of the jaws was captured with a high resolution digital camera. The image captured as shown in Figure 5-7 was transferred to a personal computer. The jaws’ tip displacement was then determined by converting the number of digital pixel in the image into real geometry in mili-meters.

Driving Circuit Power Measurement

The aim of this test was to determine the efficiency of the driving circuit. This was done by probing the input of the circuit with a multi-meter, measuring the input current as well as the input voltage as shown in Figure 5-8.

Similarly, at the output of the circuit, the average output current (SMA excitation current) was measured. With the resistance of the load (SMA wire) known, the input power and the output power of the driving circuit can then be calculated. Subsequently, the efficiency of the driving circuit was determined.

Chapter Six

RESULTS AND DISCUSSION

This chapter presents results and discussions on the outcome of tests conducted on prototype gripper as discussed in Chapter 5.

Gripping Force Test Result

From the gripping force test, initial result indicated that the SMA wire of 37 mm and 50 mm diameter were too thin and fragile to act as the actuator of the gripper. Apart from being extremely prone to breaking, they are also very difficult to handle resulting in installation difficulties. As for the SMA wire of 250 mm diameter, the result was totally opposite. Its diameter was found to be too thick resulting in the requirement of higher power for activation. According to the manufacturer’s data (refer to Appendix B), the recommended excitation current for a 250 mm diameter SMA wire is about 400 mA, way beyond the capabilities of the driving circuit presented in Chapter 4, which is in the region of 200 mA to 300 mA.

With three of the five available sizes of SMA wire out of the picture, the focus of the test was shifted to the remaining sizes of 100 mm and 150 mm diameter. From the data collected as shown in Table 6-1 and subsequently presented in the chart in Figure 6-1, it was clear that the gripping force increased with the excitation current. For the 100 mm diameter SMA wire, at least 200 mA of excitation current was required for the gripper to close its jaws. The forces recorded at this current level averaged to 0.0748 N from 3 readings. With an incremental step of 25 mA for each reading, the average gripping force increased steadily to 0.1687 N, 0.2345 N and 0.3270 N respectively.

As for the 150 mm diameter SMA wire, it basically shared the same behavior with the 100 mm diameter wire. A notable difference was found to be its relatively higher gripping force at the same current level. The minimum current required for the gripper to close has increased to 250 mA, giving an average gripping force of 0.1632 N. For the given driving circuit, it was able to draw up to the maximum of 300 mA providing a gripping force of 0.5883 N.

From the collected data, a comparison was made with a theoretically computed value. From data sheets (refer to Appendix B), the SMA wire contraction force (recommended recovery force), FSMA as well as the compression spring constant, K can be obtained as followed.

From the free body diagram shown in Figure 6-2, the effective pulling force of the SMA wire, FP can be computed as followed.

With surface friction between the sliding unit and the stopping block assumed negligible, the gripper’s gripping force, FG (as shown in Figure 6-3) can be computed as followed.

As shown in the outcome of this comparison exercise, the computed gripping force (using the recommended SMA wire contraction force) was very close to the result obtained from the gripping force test. The theoretically computed force was 0.257 N while the maximum gripping force from the test was 0.327 N. This effectively proved that the gripper design utilized the SMA wire within its capability limits without unduly overheating or overstressing it. For the sake of comparison, the same set of computation done on the 150 mm diameter SMA wire gave a similar result with the computed gripping force (using the recommended contraction force, FSMA of 3.2373 N) being 0.6148 N while the maximum measured force was 0.5883 N.

Through the test, it was also found that the cycle time required for actuation (open and close) was relatively consistent for both sizes of SMA wire. As shown in Table 6-2, the time required to perform a close-open motion for the 100 mm diameter SMA wire was relatively shorter than its 150 mm counterpart. This was obvious as the smaller the cross sectional area, the faster the SMA wire absorb heat and thus the shorter time for activation required.

Gripper Cyclic Performance Test Result

In the cyclic performance test of the prototype gripper, the outcome was quite satisfactory. After being put through more than two million cycles of close-open exercise covering a period of almost two months, it can be concluded that the prototype gripper worked quite consistently throughout the test. As shown in Table 6-3, the initial few tests were aborted prematurely due to the problem of low power in batteries. The problem was immediately solved with an AC to DC adaptor as a replacement. Since then, the test was conducted virtually non-stop throughout the testing period. Only a short period of break once a while was given to allow for inspection of the gripper condition.

As shown in the chart in Figure 6-4, the target of the test was achieved with a total number of more than 2 million cycles. The final test (Test 14) was the longest and the most adventurous with a non-stop test of more than a million cycles. This single test alone took almost a month to conduct and result was favorable. Shortly after this exhaustive test has ended, the gripper was put through a brief gripping force test to gauge its gripping performance. As it turned out, the result was consistent with the result from the gripping force test discussed in section 6.1, returning an average of 0.3 N for a full powered SMA excitation. With the result of this test, it can be concluded that the reliability achieved with the SMA wire in this gripper design was creditable.

To further convince that the driving circuit being used does not overstrain the SMA wire, an approximated measurement was done on the CAD model of the gripper as shown in Figure 6-5.

The objective of this exercise was to ensure that the percentage change in the length of SMA wire does not exceed the maximum recommended level of 5 %. A simple calculation was done to approximate the length of SMA wire before and after contraction.

As it turned out, the percentage of SMA wire contraction adopted in the gripper design was way below the recommended level of 5 %. This was desired as according to the manufacturer (Mondo-Tronics) [13], the low percentage of contraction in SMA wire prevents the occurrences of overstressing and ensures that the SMA wire operates in the de-twinning region thus giving a longer lifetime.

Gripper Characterization Result

In the characterization of the prototype gripper, the results obtained are presented as followed.

Tip Displacement

In this test, the SMA excitation current was varied in steps of 25 mA starting from 0 mA to the current level where the gripper closes completely. As shown in Table 6-4, at a current level 50 mA and below, no significant displacement was observed.

With a current level of 75 mA and above, the gripper displacement-current characteristic can be modeled after a third-order polynomial function as shown in Figure 6-6.

With these readings available, the input power and output power of the driving circuit can be computed as followed. Subsequently, the efficiency of the driving circuit can then be determined.

It can be seen that the efficiency achieved in this driving circuit was rather satisfactory with a significant amount of the input power consumed by the SMA wire. This was expected due to nature of the circuit design. With sourcing current (from the 555 timer) sufficient to activate the SMA wire, no “burn up? resistor was used to protect the SMA wire from overheating. Thus most power from the output was channeled to activate the SMA wire.

However, a careful check with the MW performance table in Appendix B showed that the recommended power is 4.86 W/m which worked out to be about 0.17 W for this design. This indicates that the power dissipation of 0.3063 W found in the SMA wire may be too high. But results from the cyclic performance test showed that this large power dissipation in the SMA wire does not degrade its performance. As a precaution measure, this problem can be rectified by simply adding in a small variable resistor of 10 W in series with the SMA wire as shown in Figure 6-7. Of cause, the trade-off of doing so would caused the circuit to be less efficient with the variable resistor functioned as a “burn up? device.

Chapter Seven

CONCLUSIONS AND RECOMMENDATIONS

Conventional means of actuation in a micro-gripper mostly utilized piezoelectric actuators. These actuators are expensive and require special mounting method. Currently, there are very little micro-gripper in the market that uses shape memory alloy in particular SMA wire as a mean of actuation.

In this project, SMA wire was used to actuate an angle gripper. The gripper with a maximum jaw opening of 1 mm was designed to handle micro-components of sizes up to 500 mm. With the use of conventional milling machineries, the gripper was test fabricated and examined to evaluate its basic gripping performance. Driven by a specialized electronics circuit, the small piece of SMA wire was excited to create a pulling effect enabling the closing of the gripper jaws.

Tests conducted on the prototype gripper have shown convincing results in terms of its consistency and reliability. Its displacement and force generated does not decrease after more than 2 million cycles of alternating activation reveal that the gripper is able to support continuous long operation without any significant degrade in its performance. Though its actuating speed was not excellent, it was satisfactory given the application involved does not require lighting fast reactions.

With a pre-configured gripping force of 70 mN up to 500 mN, the gripper is ideal for applications such as micro mounting of micro-mechanical and micro-optical parts as well as the positioning of glass fiber. With a little modification in the selection of material and improve in machining accuracy, the design concept of the gripper can also be implemented in the field of micro surgery such as in MIS and the handling of bio-batteries for implantable devices.

Further development can be done to optimize the design in the aspect of the gripper’s aesthetic and ergonomic. Alternative material can also be explored with this gripper design. Some possible alternative actuators are NanoMuscle’s linear actuators [20] and Electroactive Polymer (EAP) which is also known as Contractile Polymer [21-23]. Another option would be a relatively new actuating material known as magnetic shape memory (MSM) [24] which is actuated by external magnetic field and is found to have greater actuating speed and stroke.

Furthermore, a closed-loop control on the gripper’s actuation can be implemented with suitable sensing devices in placed. Alternatively, a temperature control scheme could be developed as a replacement of the current limiting method presently being employed. Peltier devices which are also known as thermoelectric (TE) modules could be employed as an effective way to heat and cool the SMA wire [25], although by doing so, a heat-sink is required and there is the trade-off of increase in the weight and size.

References

  1. Dario, P., Carrozza, M.C., Benvenuto, A., and Menciassi, A. 2000. Micro-systems in biomedical applications. Journal of Micromechanics and Microengineering. 10: 235-244.
  2. Van Brussel, H., Peirs, J., Reynaerts, D., Delchambre, A., Reinhart, G., Roth, N., Weck, M., and Zussman, E. 2000. Assembly of microsystems. Annals of the CIRP. 49(2): 451-472.
  3. Center of Research in Microengineering, Pontedera, Italy. http://www-crim.sssup.it/research/projects/microman/default.htm. Visited on 22nd August 2004.
  4. Speich, J., and Goldfarb, M. 2000. A compliant-mechanism-based three degree-of-freedom manipulator for small-scale manipulation. Robotica (2000). 18: 95-104.
  5. Roch, I., Bidaud, Ph., Collard, D., and Buchaillot, L. 2003. Fabrication and characterization of an su-8 gripper actuated by a shape memory alloy thin film. Journal of Micromechanics and Microengineering. 13: 330-336.
  6. Krulevitch, P., Lee, A.P., Ramsey, P.B., Trevino, J.C., Hamilton, J., and Northrup, M.A. 1996. Thin film shape memory alloy microactuators. Journal of Microelectromechanical Systems. 5(4): 270-282.
  7. Morra, F., Molfino, R.M., and Cepolina, F. 2004. Miniature gripping device. Paper presented at the IEEE International Conference on Intelligent Manipulation and Grasping IMG 04, 1-2 July 2004, Genova, Italy. 363-368.
  8. Sallé, D., Cepolina, F., and Bidaud, P. 2004. Surgery grippers for minimally invasive heart surgery. Paper presented at the IEEE International Conference on Intelligent Manipulation and Grasping IMG 04, 1-2 July 2004, Genova, Italy. 172-179.
  9. Ku, S., and Salcudean, S.E. 1996. Design and control of a teleoperated microgripper for microsurgery. Paper presented at the IEEE International Conference on Robotics and Automation ICRA 96, 22-28 April 1996, Minneapolis, Minnesota. 1: 889-894.
  10. Tokhi, M.O., and Azad, A.K.M. 1997. Design and development of an experimental flexible manipulator system. Robotica (1997). 15: 283-292.
  11. Mondo-tronics, Inc. Electric Piston Instructions. http://www.mondo-tronics.com/download/EPiston_Instr_V12.pdf. Visited on 5 September 2004.
  12. Bertetto, A.M., and Ruggiu, M. 2003. A two degree of freedom gripper actuated by sma with flexure hinges. Journal of Robotic System. 20(11): 649-657.
  13. Gilbertson, R.G. 2000. Muscle wire project book 3rd ed. San Rafael, CA: Mondo-tronics. 1-4 to 3-22.
  14. Mosley, Michael J., and Mavroidis, Constantinos. 2001. Experimental nonlinear dynamics of a shape memory alloy wire bundle actuator. Journal of Dynamic System, Measurement, and Control. 123: 103-105.
  15. Stalmans, R. and Van Humbeeck, J. 1995. Shape memory alloys: functional and smart. Paper presented at the Smart Materials and Technologies – Sensors, Control Systems and Regulators Seminar, October 1995, Prague, Czech Republic.
  16. Peirs, J., Reynaerts, D., and Van Brussel, H. 2001. The ‘true’ power of sma micro-actuation. Paper presented at the Micromechanics Europe Workshop MME’01, 16-18 September 2001, Cork, Ireland. 217-220.
  17. Kohl, M., Dittmann, D., Quandt, E., and Winzek, B. 2000. Thin film shape memory microvalves with adjustable operation temperature. Sensors and Actuators. 83: 214-219.
  18. Ma, N., and Song, G. 2003. Control of shape memory alloy actuator using pulse width modulation. Smart Materials and Structures. 12: 712-719.
  19. Tekscan Inc. FlexiForce. http://www.tekscan.com/flexiforce/flexiforce.html. Visited on 15 January 2005.
  20. NanoMuscle Inc. NanoMuscle Linear Actuator. http://www.nanomuscle.com/. Visited on 5 September 2004.
  21. Del Cura, V.O., Cunha, F.L., Aguiar, M.L., and Cliquet, A.Jr. 2003. Study of the different types of actuators and mechanisms for upper limb prostheses. Artificial Organs. 27(6): 507-516
  22. Bar-Cohen, Y., and Leary, S. 2000. Electroative polymers as artificial muscles changing robotics paradigms. Paper presented at the National Space and Missile Materials Symposium, 27 February – 2 March 2000, San Diego, CA.
  23. Monkman, G.J. 2000. Advances in shape memory polymer actuation. Mechatronics. 10: 489-498.
  24. Suorsa, I., Pagounis, E., and Ullakko, K. 2004. Magnetic shape memory actuator performance. Journal of Magnetism and Magnetic Materials. 272-276: 2029-2030.
  25. Potapov, P.L. 1998. Thermoelectric triggering of phase transformations in Ni–Ti shape memory alloy. Materials Science and Engineering. B52:195-201.
  26. RS Components Pte. Ltd. Compression Springs (821-267). http://www.rssingapore.com/. Visited on 10 December 2004.

Please be aware that the free essay that you were just reading was not written by us. This essay, and all of the others available to view on the website, were provided to us by students in exchange for services that we offer. This relationship helps our students to get an even better deal while also contributing to the biggest free essay resource in the UK!