I took this on as an interactive art project to get acclimated with working with the shape-memory alloy nitinol wire and look at some of its properties. I think the result is pretty cool.
So here’s what I built- a “man-doing-a-pullup” module comprised of two lengths of muscle wire in series, connected by a rigid conducting bar to allow the body to stay naturally perpendicular to the ground, as well as for the circuit to close when I plug in the terminals- which I fit in a 3d-printed housing I designed. I also have a bit of solder and copper tape to stick everything to the paper “Mister Muscle(wire)”, which I illustrated to give a bit of character.
How’d I make it? Well, it took a couple of tries. After trying with no success with lower voltage batteries, burning through some structural stitching, and troubleshooting undesired ‘memorized’ states, I was finally able to get the wires to work for me. I ended up pre-training the wire by clamping them into the bent shape with pliers, and training them to shape with my soldering iron on high. Later on, I was able to achieve finer adjustments by running a 9v battery through and forcing it into shape with pliers.
What would I do next? I want to try for an entirely autonomous system- the original idea for this was using a 555 timer, as well as a pair of antagonist muscles, to time a full pull-relax cycle. Another idea was to make an even more self-contained module that didn’t even need the scaffolding, and has magnetic hands to grab onto metal objects to complete the circuit, and have it start doing pullups. Two issues I ran into here were that the magnets I was using were becoming demagnetized at high temperatures, and the original battery I had in mind didn’t have a high enough voltage to heat the wires to the necessary temperatures.
Maybe I’ll add these features in when I revisit the project in the future.