Moment of Inertia, Issue #1 | The Future of Failure

Welcome to the first edition of the latest member of the Magnitude and Direction family.

Moment of inertia is a place where I'll be giving more of my commentary, in a longer form, on select topics I cover in my newsletter Magnitude and Direction. The goal is to give you 10 to 15 minutes of interesting reading on a topic while you're still in bed - your moment when the inertia of being cozy in bed keeps you there, even though you're awake and getting ready for the weekend morning. I hope you enjoy hearing what I have to say about the topics I cover in M&D and encourage you to participate in the discussion as well.

For this first edition, we're talking about something very close to my heart: 3D Printing.



The Future of Failure.

3D printing comes in many shapes and sizes, and utilizes a variety of different materials to produce objects literally from the ground up. One thing that's consistent across all the different 3D printing modalities, however, is the need to have a foundation that each printer layer can sit upon.

In the case of the first layer, this is the build platform where the initial material is deposited. For all subsequent layers, however, it falls to the previously printed material to support the new layers added by the printer. Without material to provide a foundation, the 3D printing process quickly fails as material is deposited in unintended places.

Nowhere is this phenomenon more apparent than Fused Deposition Modeling, or FDM, the 3D printing process most people are familiar with. In this approach, plastic is heated up and extruded from a nozzle while being moved in 3D Cartesian space. As the print head moves, plastic is deposited and quickly cools, building up the printed structure. 

A close-up of the FDM printing process.

A close-up of the FDM printing process.

When there isn't an existing layer of plastic present to support the new printed layer, however, the results begin to look more like strange spaghetti than a functional object:

When there’s no material to support a new layer, the results can be quite… stringy.

When there’s no material to support a new layer, the results can be quite… stringy.

An outcome like the one above is often considered a failure. However, as the Filament Sculptures article in yesterday's edition of Magnitude and Direction shows, even an erroneous process can be exploited in a controlled way. One person's trash is another one's treasure, and one person's failure is another person's art.

But art is just the beginning. There are all kinds of ways we can exploit the unintended consequences of manufacturing processes to create things that we either couldn't before, or would have a very difficult time doing with other, specialized equipment. I will say that, on this front, the arts community is the vanguard, leaving the manufacturing community far behind. It's understandable why this would be the case. In the manufacturing world, you don't typically have the time to explore all the things you can get your machines to do when you use them in unexpected ways - products need to be produced and shipped out the door. If a printer isn't producing a clean, smooth finished part, then it's not working correctly. In the arts, on the other hand - especially the digital arts - the whole point is to push the boundaries of technology and find out what happens when we use tools in unexpected ways.

This is something the hardware and manufacturing communities should try to engage in more often. While 9 out of 10 times the results won't yield anything useful, the 10th experiment can often produce a result that more than compensates for the failed first nine.

This happened to me recently in the course of my PhD studies. As part of my work, I need to produce very small, detailed, and geometrically intricate structures, with the end goal of incorporating these novel structures into medial device prototypes. There's one big problem, though: these structures are too small to make using standard fabrication techniques like 3D printing or CNC machining, and too big to produce using nanofabrication techniques like lithography and molecular beam epitaxy.

What I was able to do, however, was 3D print the negatives, or opposites, of the structures I ultimately wanted and then cast various rubbers into those molds in order to produce the final object I needed. This produced its own set of problems, however, as more often than not I ended up destroying the part as I tried to remove it from the mold. I needed a way to produce a rubber mold that I could "melt" away.

As it so happened, the 3D printer I'm using to produce these molds prints in two materials: a hard plastic and a softer wax, which is used only as support, to provide a foundation for layers of the hard plastic that hang to far out to be layers on top of earlier layers of hard plastic. Just last week, it occurred to me that I might be able to "hack" this system so that only the wax portion got printed, and in the shape of the molds I needed.

The idea rests upon understanding how the printer is designed to operate and then pushing the limits of that normal operating procedure. Any portion of a model that is free-floating, or not resting upon a previous layer is supported by a foundation of wax. A solid cube will not need any wax, but a hollow cube (with the bottom face removed, leaving the interior open) will.

A solid cube is always able to support the next layer upon the one that came before it, but a hollow cube needs additional support wax to hold up its “roof”.

A solid cube is always able to support the next layer upon the one that came before it, but a hollow cube needs additional support wax to hold up its “roof”.

It turns out that, if you make the walls of the hollow cube thin enough, the printer isn't actually capable of printing them - they're too thin. It is still capable of understanding that those walls exist mathematically, though, and, as such, will print wax to support them, even if it can't actually produce them. The result is a printed wax form that exactly matches the dimensions of the original model. The difference is that, instead of being built from a hard, brittle plastic, it's made out of a soft, meltable wax.

All of a sudden, I had my ephemeral rubber molds, that I could pour material into and then melt away after the rubber had solidified. Furthermore, this technique has enabled me to produce even more complicated geometries - forms I would never have had a chance of separating from my cast rubber previously. Now, though, all I have to do is heat up the mold once the rubber is cured and the material falls away.

There are numerous potential applications for this new fabrication approach, and I'm working right now to fully characterize it and share it with the community - maybe even publish a short paper on it. This major personal breakthrough, however, wouldn't have been possible if I hadn't tried to push the 3D printer to a place where it wasn't supposed to work properly. Don't forget, that machine isn't supposed to print in only wax.

The takeaway message here is that we should pay as much attention to our failures and negative results as we do our successes and positive outcomes. You never know when you might be able to make lemonade from lemons...