CECS Spotlight: Smart stuff

'Smart materials' are being exploited by ANU systems engineers who believe they can tame their shape recovery properties to develop an actuator that could become a fast, accurate and lightweight component in electronics, toys and medical appliances.

The pair developed a pointer-like device, driven by two of their special actuators, and commanded it to trace the outline of a square. Actuators are a component of mechanical devices that control or move a system - such as that which allows cardiologists to change the direction of an endoscope inside the body.

The researchers from the ANU College of Engineering and Computer Science hope that their actuator will quickly and accurately direct the pointer to move around the outline of the square. But theirs is no ordinary actuator, as it is driven by wires made of Shape Memory Alloys (SMAs). An SMA has properties that make it different from any ordinary alloy: when an SMA is cool it can be easily deformed, but when heated it transforms back to its original shape. This chameleon attribute is thanks to a molecular rearrangement in the alloy during heating and cooling, known as a phase transformation, which generates forces and produces motion. "The bigger the current, the faster the heating, and so faster motion is achieved," Featherstone says.

But because the 'shape recovery properties' are generally slow and difficult to control, SMA wires have yet to become a common material in actuators. Making SMA wires more accurate at higher speeds is the challenge Featherstone and Teh have taken up with gusto.

In the tracing experiment, the control system of the SMA wires inspires the pointer to trace the outline of the square slowly, but with reasonable accuracy - it handles the right angles of the square well and does not undershoot or overshoot along the sides of the square. But as the speed of the pointer picks up, larger currents are applied to the wires, and it loses accuracy.

"Eventually it gets to be so fast it's missing the corners," Featherstone says. "At this point you've got speed, but you don't really have accuracy."

But the two researchers have already figured out a solution to the problem of large currents overheating the SMA wires and damaging the actuator. They've called their breakthrough the 'Rapid Heating Algorithm' and are in the process of patenting it.

"We've invented a simple control method to combine rapid heating with safety measures to prevent overheating. It involves measuring the electrical resistance of the SMA wire, and exploiting the way that its resistance changes with temperature. This can be done using cheap, compact electronic components," Featherstone says.

"It allows the SMA actuator speed to increase, by up to a factor of two, compared to existing control schemes. Faster speed means reduced response time, and faster completion of the task at hand. So it represents quite a step forward."

SMA wires are also strong. If they encounter any resistance during the phase transformation, large forces are generated. A wire with a diameter of one millimetre can lift a weight of 15 kilograms, Feathersonte says. This is the basis for the actuating mechanism.

Fast and accurate SMA actuators would be more compact, mechanically simple and light. Unlike electric motors, SMAs can operate quietly and cleanly, would create no sparks, and require no gears or motor brushes. They don't shed lubricants or dust, and nickel-titanium SMA is bio-compatible, which means it could be used inside living tissues.

Having solved the problem of overheating the SMA wire, the researchers are now focused on improving the motion control - the overshooting and undershooting of the pointer's motion. Featherstone describes the process as being like tweaking a recipe over and over again to get the perfect result. "We make improvements after improvements but very incrementally," he says.

SMA actuators are already used in things like medical devices (such as endoscopes, steerable catheters), electronic latches and switches, where they have to be accurate but not speedy. But a fast and accurate version able to deal with heavy loads would open up new innovations in electronics, robotics and small consumer technologies, such as smaller digital cameras.

The shape recovery properties of SMAs were first discovered in the 1930s, but it wasn't until the 1960s that the potential applications of this attribute came to be understood.

Featherstone's interest in SMAs and their potential as actuators developed later still, after he read the PhD thesis of a Canadian colleague's student, Danny Grant, in 1999.

"The thesis, put simply, focused on coiling these SMA wires in an accordion arrangement to amplify the strain that could be produced. But he also did some experiments on force control and published some graphs. I happened to notice that these graphs showed that the SMA wires had an almost instant response to the heating. Twenty milliseconds later there's already a force response from the wire," Featherstone says.

"What was curious was that it's conventional wisdom that SMAs are slow, yet he had demonstrated that they were really quite fast for small motions. It is only the large motions that are slow."

Featherstone shelved the thesis for a few years, but the findings stayed in the back of his mind. When Yee Han Teh arrived at the research school to take up a summer research scholarship at the end of 2002, developing an SMA actuator presented a good project for an enthusiastic scholar. In the end, advances on the actuating properties of SMAs became Teh's honour project, and are now his PhD thesis topic.

"From these graphs in Grant's thesis, we figured that actually SMA actuators could move really quickly - but just in small amounts. The idea is that if we could control the force really quickly, we can incorporate that in the position control of practical SMA actuator applications to achieve fast and accurate motion. That was really the inspiration for this project," Featherstone says.

Though there was the theory that this approach could work and would be stable, transferring the theory to practice was assigned to Teh. He spent his early PhD designing and building a new experimental test bed, which allowed him to conduct modelling experiments on which he could tweak all the inputs - stress, strain and current - to obtain valuable information about the wires. He did hundreds and hundreds of tests.

"Although it's a strong material, we have gone through a lot of wires," Teh says. "Just apply a little excess force on, or too much heating current through, and the balance is tipped, and snap."

"We have these little lights on the latest test bed," Featherstone explains. "Green is 'okay', orange is 'oh dear', red is 'broken'. They indicate load and if it goes beyond a certain level then it's broken - the red light stays on." These alert signals are an additional safety feature to Teh's second, more highly resolved, test bed. Safety features such as these prevent damage to the experimental test bed, which took a year to design, build and commission.

"Even in larger industrial applications, simulations are quite important, because you don't want to risk equipment damage or injury," Featherstone says. "The modelling helps us to decipher, control and tune up the system, and once that's done we can test it on the test bed and fine tune it further to get the best performance for a control system."

At the top of the test bed, there is a linear motion stage that moves a pulley up and down. The pulley has a short chord attached which ends in two small eyelets for the SMA wires to pull on antagonistically. At the bottom, there is a pair of sensitive load cells, which measures the force on the SMA wires individually.

A load can be attached to the pulley, which enables the researchers to test the relationship between external load, speed and accuracy. The wires themselves are 80 centimetres long, but are doubled-up so that the two ends are connected at the load cell and the middle passes through the eyelets at the top.

The linear stage can generate motions with an accuracy of one micrometer, and the load cells can measure forces with a resolution of 0.3 milli-Newton. The test bed electronics can also precisely deliver heating currents using the rapid heating algorithm and send all the data collected to a computer for analysis.

According to Featherstone, this test bed is matchless in its resolution and experimental capacity. "There are quite a few labs that are trying to do something with SMAs, but there's nobody doing it like us."

Teh also believes the work is singularly important. "From my perspective, ours is more of an enabling technology," he says. "Other groups are doing similar work, but it's more like they are doing specific applications-based research, such as designing an SMA actuator for endoscopes. Our work aims to refine a good control system that will apply to any actuator application."

They are quietly confident of having a highly resolved motion control system using SMA wire actuators ready for patenting soon.

They say areas where their SMA actuator might prove a success include automotive applications such as thermal sensors, positioners, mirror tilting mechanisms; appliances like camera anti-shake compensators, temperature cut-off mechanisms, micro-grippers, micro-positioners; and robotics in commercial and industrial settings.

"There are many areas, possibly the biggest potential is in mechanical instrumentation, where one SMA actuator can just be plugged into an application and do all the work," Teh says.

"There are also possibilities in small consumer products," Featherstone says. "An example might be in a tiny pan-tilt mechanism for a very, very small camera. SMA wires are light and are most efficient at short lengths, so this would be an ideal application. Basically we think it can be applied to anything where a change of direction is needed quickly, that also needs to be small, light and cheap."

More information:http://users.rsise.anu.edu.au/~roy/SMA/