Muscle.Wire

Muscle Wire is an extremely thin wire made from nitinol (a nickel-titanium alloy) that is known for its ability to contract when a voltage is applied. This change in state is produced by the heat from the current - which causes the wire to take on a hardened form. This physically shortens the wire, and when the voltage is removed, the wire expands back to its normal length and shape as it cools.

Muscle Wire is excellent for creating small movement in tight areas. Because it is so thin, it can fit in most places and can bend up to 90 degrees without contraction problems (and risk of breaking). In most cases, they don't use too much power, so a simple battery pack usually works well. Yet, as cool as Muscle Wire is, it does have it's problems.

Because the wire (also referred to as flexinol muscle wire) basically produces its heat through its own resistance, it's very easy to burn out. The amperage must be pretty close to exact for the process to actually work, and figuring out the proper amperage is usually a hassle - based upon the diameter of the wire you work with, the length of the segment of wire, the resistance of that wire, and the percentage of shrinkage you desire. Muscle Wire is also difficult to work with because the current must be closely regulated. If current is applied for too long of a time, the wire burns out. It is therefore necessary to build either an oscillator circuit or program your microprocessor to shut off the current after about 2 seconds (which will contract the wire and then allow it to relax). The last thing to watch out for when working with flexinol are the connections. Because the wires are so painfully thin (think: human hair) they are sometimes difficult to work with, and all connections must be crimped.

Cost: ~$10 per meter

Marionette Project: Afraid of Flying
Initially, I set out to use the wire to control a human figure created out of wire and plaster. Wanting to move away from the marionette model of vertical strings manipulating body parts, I began using small lengths of the wire to function as mini-muscles - pulling in order to contract a joint and relaxing to allow the joint to fall back into an open position. After calculating the appropriate current for the 2cm length (and after burning out quite a few wires in the process) I realized that 5% contraction of a 2cm wire isn't very much at all - hardly enough to even jerk the marionette. Concurrent with this eureka moment, the piece of muscle wire burnt out. How appropriate.

Flustered and frustrated, I began to flip further into the muscle wire project book - hoping for inspiration. After finishing the book, one thing was clear - muscle wire was tough to work with. Hardly explaining the importance of the oscillation circuitry or the math involved to derive the correct currents, the workbook settled into using only 10cm lengths of wire (luckily the same diameter that I had purchased) and a method of powering the wire that consisted of these instructions - "Now touch the battery to the wire...NOW QUICKLY REMOVE THE VOLTAGE BEFORE YOU BURN OUT THE WIRE." It was unsettling at best.

However, I had finished the book. I had (supposedly) learned all about muscle wire. What I really learned was their trick: always use 10cm lengths of muscle wire. 10 cm works surprisingly well because you can use 3 volts (ie: two AA batteries) to power the wire. This was a godsend. The book also took me back to high-school physics class and reintroduced me to the great powers of the lever. Using levers, I was able to maximize the 5% shrink factor of the wire and convert this energy into a much larger movement. This technique allowed me to use a series of four levers to control 5 strings attached to the marionette. To give the figure a bit more personality, I attached one of the wires to his hand, one to an elbow, two to his knees, and one to his tailbone. This configuration allowed for a variety of movements and a specific contraction into the fetal position that was essential to the narrative.

At one point, I was going to connect a cell-phone vibrator to the torso of the marionette to provide a way for him to shake ("cry")... but after reading about Vardit's experience with burning out multiple 3 volt cell-phone vibrators during her project for last year's show, I left my vibrator in the package and continued on with the more important aspects of the project.

Setting up the breadboard and programing the bx-24 microprocessor to handle the information exchange between the director controlled hypertext and the marionette proved a little challenging. Nothing too difficult - it just took a lot of brain power to figure out how to time the pulses of voltage being sent to the muscle wire. Initially, I attempted to send the 3 volts with the chip - using the 2 AA batteries to power the entire chip. This didn't work at all - the little light just kept turning on and off on my chip. I assumed that this would work fine and that I just had wires crossed...so I spent LOTS of time trying to figure that out. Unfortunately, I had missed the fine print in the bx-24 handbook that says not to send less than 4.7 volts into the microprocessor (or a continual resetting will occur). This wasn't too major, and eventually it was corrected using reed relays to send the alternative 3 volts to the muscle wires.

The two second delay was created in the programming. When a text link was clicked in director, a single variable (referencing a specific movement of the muscle wire) was sent to the chip. This turned a muscle wire on, ran a delay of 2 seconds, and then set all of the muscle wires back to 0 voltage (just as a fail-safe device).

Thinking I was ready for anything, I plugged it all together and set out to use a multimeter to test output voltages before putting that 3 volts anywhere near the easily fried muscle wire. Yet again...problems. This was the most frustrating because I had planned so well (and for so long). Nothing worked. I debugged. I debugged again. I rewrote. I debugged again. And alas...it was something simple. I hadn't hooked up my serial cord correctly (chalk it up to a lack of sleep) and that was it. An easily correctable human error (and NO...I don't feel bad for thinking it was my bx-24 or blaming my chip for anything).

Finally, it works. The hypertext controls the marionette's movement and the muscle wires have continued to work well with the 2 second on-cycle. Watching the project in action, I have mixed feelings as to whether or not the wire works to the advantage of my project. It's contraction rate is pretty slow, which creates an even fluid motion in the marionette - which is nice, but sometimes, speed is nice as well. It does provide a dreamy effect which plays into the memory space created in the director piece. Yet, there's something that doesn't quite work with it. There's something absolutely horrible about having part of the project screen based and part in the physical world. Focus is split, and the marionette almost feels like it is TOO controlled by the computer. The whole experience is best in front of a projector where the marionette becomes part of the world of the story - rather than just being a physicalized part of the whole

If anyone is considering working with muscle wire, I would recommend that they begin their project early so that they have enough time to master the wire before even introducing it into their project environment.




Muscle Wire is an extremely thin wire made from nitinol (a nickel-titanium alloy) that is known for its ability to contract when a voltage is applied. This change in state is produced by the heat from the current - which causes the wire to take on a hardened form. This physically shortens the wire, and when the voltage is removed, the wire expands back to its normal length and shape as it cools. Muscle Wire is excellent for creating small movement in tight areas. Because it is so thin, it can fit in most places and can bend up to 90 degrees without contraction problems (and risk of breaking). In most cases, they don't use too much power, so a simple battery pack usually works well. Yet, as cool as Muscle Wire is, it does have it's problems. Because the wire (also referred to as flexinol muscle wire) basically produces its heat through its own resistance, it's very easy to burn out. The amperage must be pretty close to exact for the process to actually work, and figuring out the proper amperage is usually a hassle - based upon the diameter of the wire you work with, the length of the segment of wire, the resistance of that wire, and the percentage of shrinkage you desire. Muscle Wire is also difficult to work with because the current must be closely regulated. If current is applied for too long of a time, the wire burns out. It is therefore necessary to build either an oscillator circuit or program your microprocessor to shut off the current after about 2 seconds (which will contract the wire and then allow it to relax). The last thing to watch out for when working with flexinol are the connections. Because the wires are so painfully thin (think: human hair) they are sometimes difficult to work with, and all connections must be crimped. Cost: ~$10 per meter Marionette Project: Afraid of Flying Initially, I set out to use the wire to control a human figure created out of wire and plaster. Wanting to move away from the marionette model of vertical strings manipulating body parts, I began using small lengths of the wire to function as mini-muscles - pulling in order to contract a joint and relaxing to allow the joint to fall back into an open position. After calculating the appropriate current for the 2cm length (and after burning out quite a few wires in the process) I realized that 5% contraction of a 2cm wire isn't very much at all - hardly enough to even jerk the marionette. Concurrent with this eureka moment, the piece of muscle wire burnt out. How appropriate. Flustered and frustrated, I began to flip further into the muscle wire project book - hoping for inspiration. After finishing the book, one thing was clear - muscle wire was tough to work with. Hardly explaining the importance of the oscillation circuitry or the math involved to derive the correct currents, the workbook settled into using only 10cm lengths of wire (luckily the same diameter that I had purchased) and a method of powering the wire that consisted of these instructions - "Now touch the battery to the wire...NOW QUICKLY REMOVE THE VOLTAGE BEFORE YOU BURN OUT THE WIRE." It was unsettling at best. However, I had finished the book. I had (supposedly) learned all about muscle wire. What I really learned was their trick: always use 10cm lengths of muscle wire. 10 cm works surprisingly well because you can use 3 volts (ie: two AA batteries) to power the wire. This was a godsend. The book also took me back to high-school physics class and reintroduced me to the great powers of the lever. Using levers, I was able to maximize the 5% shrink factor of the wire and convert this energy into a much larger movement. This technique allowed me to use a series of four levers to control 5 strings attached to the marionette. To give the figure a bit more personality, I attached one of the wires to his hand, one to an elbow, two to his knees, and one to his tailbone. This configuration allowed for a variety of movements and a specific contraction into the fetal position that was essential to the narrative. At one point, I was going to connect a cell-phone vibrator to the torso of the marionette to provide a way for him to shake ("cry")... but after reading about Vardit's experience with burning out multiple 3 volt cell-phone vibrators during her project for last year's show, I left my vibrator in the package and continued on with the more important aspects of the project. Setting up the breadboard and programing the bx-24 microprocessor to handle the information exchange between the director controlled hypertext and the marionette proved a little challenging. Nothing too difficult - it just took a lot of brain power to figure out how to time the pulses of voltage being sent to the muscle wire. Initially, I attempted to send the 3 volts with the chip - using the 2 AA batteries to power the entire chip. This didn't work at all - the little light just kept turning on and off on my chip. I assumed that this would work fine and that I just had wires crossed...so I spent LOTS of time trying to figure that out. Unfortunately, I had missed the fine print in the bx-24 handbook that says not to send less than 4.7 volts into the microprocessor (or a continual resetting will occur). This wasn't too major, and eventually it was corrected using reed relays to send the alternative 3 volts to the muscle wires. The two second delay was created in the programming. When a text link was clicked in director, a single variable (referencing a specific movement of the muscle wire) was sent to the chip. This turned a muscle wire on, ran a delay of 2 seconds, and then set all of the muscle wires back to 0 voltage (just as a fail-safe device). Thinking I was ready for anything, I plugged it all together and set out to use a multimeter to test output voltages before putting that 3 volts anywhere near the easily fried muscle wire. Yet again...problems. This was the most frustrating because I had planned so well (and for so long). Nothing worked. I debugged. I debugged again. I rewrote. I debugged again. And alas...it was something simple. I hadn't hooked up my serial cord correctly (chalk it up to a lack of sleep) and that was it. An easily correctable human error (and NO...I don't feel bad for thinking it was my bx-24 or blaming my chip for anything). Finally, it works. The hypertext controls the marionette's movement and the muscle wires have continued to work well with the 2 second on-cycle. Watching the project in action, I have mixed feelings as to whether or not the wire works to the advantage of my project. It's contraction rate is pretty slow, which creates an even fluid motion in the marionette - which is nice, but sometimes, speed is nice as well. It does provide a dreamy effect which plays into the memory space created in the director piece. Yet, there's something that doesn't quite work with it. There's something absolutely horrible about having part of the project screen based and part in the physical world. Focus is split, and the marionette almost feels like it is TOO controlled by the computer. The whole experience is best in front of a projector where the marionette becomes part of the world of the story - rather than just being a physicalized part of the whole If anyone is considering working with muscle wire, I would recommend that they begin their project early so that they have enough time to master the wire before even introducing it into their project environment.

The sleeping project celebrates missed connections in people's lives. In the installation, users interact through touch with life-size video images of a sleeping body projected onto a mattress. The video of the sleeper responds to this touch and constructs a dancelike narrative in the 2-D dreamscape of the mattress. Here, we are able to interact with a sleeping virtual presence - an action not normally performed even on another human being. What does this doorway provide us access to? Can a connection ever be made? In searching for embodiment of a virtual presence, we are attempting to wake the sleeping dreamer - who once woken, ceases to exist...