February 15, 2012      

Caenorhabditis elegans is a remarkable worm. Less than one milimeter in size, the hermaphroditic, soil-dwelling nematode bears all the physiological systems we associate with higher animals including digestive, muscular, and reproductive systems.

Completely transparent, every C. elegans cell can be viewed under a microscope. This fact, allied with the simplicity of the worm’s design, has meant that C. elegans has become a model organism in both genetics and systems biology.

Locomotion is controlled by a nervous system with just 302 neurons. Nevertheless, C. elegans is incredibly agile and effective moving through cluttered spaces and displays distinct locomotive styles depending on whether it finds itself in wet or dry environments.

Inspired by the worm’s economical locomotive prowess, Jordan H. Boyle, a post-doctoral research fellow in the schools of Computing and Mechanical Engineering at the University of Leeds, created the remarkable WormBot.

Using a virtual neural circuit (which is in turn supported by an integrated neuro-mechanical model), Boyle’s 6.56 foot-long creation is now at the advanced prototyping stage and could one day be used to assist search and rescue operations by wriggling its way through debris to find survivors.

Fusing plastic, metal, ceramic or glass powders into complex 3D shapes.

There’s an unlikely hero behind the success of the WormBot project however ?the Vanguard HS (High Speed) selective laser sintering (SLS) device developed by 3D Systems Inc. SLS is a process that fuses plastic, metal, ceramic, and/or glass powders using a high-powered laser; enabling the rapid creation of complex 3D shapes.

?I have often said to people who asked me about my work, that I really don’t think the WormBot project would have been as successful as it was, if it wasn’t for that [SLS] technology. I love the technology. It just makes life so much easier. Especially for the sort of stuff I was doing –very early-stage, exploratory, robotics research,? says Boyle.

The clincher for Boyle is the speed of 3D printing technologies compared to traditional machining solutions.

At the start of the WormBot project, Boyle requested quotes from both traditional and 3D prototyping facilities. Whereas traditional CNC milling would have taken about three weeks to create parts for the 6.56 feet-long robot, the 3D printing facility could deliver the parts within 2-3 days after receipt of a SolidWorks-created 3D graphics file.

?Using CNC milling to make the sort of relatively complex shapes that went into my robot would have required quite a lot of different types of tooling and the parts would have had to have been manually shifted between different machines. Even if I had had a dedicated team of technicians, CNC milling would have taken longer than it took to print,? says Boyle.

Another reason Boyle chose an SLS device was because the mechanical properties of the finished nylon. To mount an 8 millimeter bearing, he instructed the machine to create holes 7.9 millimeters in diameter and then simply press-fitted the bearings into the robot. Similarly screws could be fitted by self-tapping, rather than going through the process of designing a screw in a 3D graphics package.

SLS is one of many additive manufacturing and rapid prototyping techniques covered in the catch-all term ‘3D printing.’ SLS-based printers build up parts in layers by fusing powders ?in Boyle’s case nylon powder– with a laser. SLS-based printers are so-called ‘consolidators’ and, at the end of printing process, the printed object is dusted to remove excess powder, revealing the finished product.

Extruder-type 3D printers, on the other hand, work by releasing a liquid (usually a plastic) from one or more nozzles and building up the object layer by layer as the liquid solidifies.

SLS-based printers are more expensive than other 3D printer types, says Boyle, but they offer three main advantages over their more-commonly available extruder-based counterparts: they provide higher resolution (in the region of .5 mm), they don’t require inner support structures which need to be broken off at the end of the printing process, and they can print with a wider range of materials, from thermoplastics and glass, through to metals.

The downside of SLS-based 3D printers for Boyle is cost.

?The parts I had made for my robot were not exceptionally cheap. If I’d been paying those costs out of my own pocket I would have been broke by now. More traditional 3D printer types are typically much cheaper,? says Boyle, who estimates that the total cost of all his 3D printed parts was just over USD$3,000.

As desktop-sized 3D printers improve and continue to fall in price, they will become ?the answer to peoples’ prayers? when it comes to low-cost, high-speed prototyping and possibly even small-scale robot production, says Boyle, but the transition to large-scale manufacturing is not likely to be so straightforward.

?In my experience, the finished product you get from a 3D printer has a distinct layered look about it and there sometimes can be a little bit of weakness around where layers join. For large-scale manufacturing, one would have to look very carefully at the manufacturing volume and cost ratios beforehand to make sure that its cheaper than more conventional manufacturing technologies.?

Broadly speaking, Boyle advises roboticists considering 3D printing to ?go for it, you’ll be glad you did? but with one caveat: ?don’t forget common sense.?

?One should definitely consider adding 3D printing to the set of tools you use, because almost any robotics project is going to benefit from at least some 3D printed parts. However, if what your robot needs is a flat chassis to which you can bolt some motors and batteries, then you’re better off cutting it from a sheet of plastic/wood/carbon fibre and saving yourself a lot of money.?

The hype surrounding 3D printing is growing ?recent articles in The Economist, Computer World, and IEEE Spectrum, have gone beyond highlighting the cost-effectiveness, speed, and prototyping capabilities of 3D printers to, in some cases, claiming that 3D printing is on the verge of becoming a mainstream tool ?not just for commercial manufacturing operations, but for consumers too.

Some of that hype is justified by rapid growth in the 3D printing market. The global market for all 3D printer products and services grew by 24 percent in 2010 to USD$1.325 billion and 2011 growth is expected to be in double digits, says Terry Wohlers, president of 3D printing market analyst firm Wohlers Associates.

Factors impeding widespread adoption of 3D printers amongst roboticists are cultural and technical, says Wohlers.

?Tradition is the most important factor. It [3D printing] requires some new ways of thinking within a company, and a willingness to try a new way of prototyping and manufacturing,? says Wohlers.

Metal-based 3D printing has only really started taking off in the past ten years, says Wohlers. And while orthopedic manufacturers have used metal-based 3D printers for a few years, their use has yet to become widespread. The dental and aviation industries are are driving demand for greater metal-printing capabilities.

Size is another factor, with the biggest printable parts typically being around 300 millimeters in one direction, says Wohlers, who expects the capacity of 3D printers to increase rapidly over the next few years, driven by demand from the aviation sector.

Finally, 3D printers can only print objects in a limited range of materials and only the very latest, largely experimental machines can print electronics. That could be a key turning point for roboticists as it would allow electronic circuits and batteries to be printed alongside the rest of the robot.

Most of the leading metal-printing 3D systems come from Europe ?a legacy of the investment made by places like Germany’s Fraunhofer Institute in 3D printing research, says Wohlers. Despite this, the largest market for 3D printing products and services is the United States.

For prototyping purposes, the speed and flexibility of even entry-level (circa. USD$15,000) 3D printers makes their use a ?no-brainer,? but for more complex prototyping companies will probably need to invest in a higher-end machine, which could cost anywhere up to USD$300K.

3D printing best suited for small volumes and small, complex parts.

Large-scale robot production is some way off, says Wohlers, but as 3D printers become capable of printing larger parts from different sorts of materials, the case for using them in full-scale manufacturing improves. For now, as far as production is concerned, 3D printing remains suited to small volumes and small, complex parts.

?I would urge robotics companies to take a look at 3D printing and run some tests, build some test parts, and evaluate the results. This is a very moving target and it’s changing quickly. It’s something that robotics companies need to monitor and watch closely, because if they don’t, their competitors certainly will and they will be at a competitive disadvantage,? says Wohlers.

To consider 3D printers simply as prototyping tools is to miss the point, says Hod Lipson, associate professor in the Creative Machines Lab (CML) at Cornell University and a pioneer of the use of 3D printing in robotics.

?There are obvious benefits in the speed with which complex structures can be made and explored. But the main benefit in the ability to explore and fabricate objects with complexities form and function that could not be manufactured in any other way,? says Lipson.

In 2010, Lipson’s team hit the headlines with the Ornithopter ?a 3D-printed, insect-like, flapping-wing robot. The 3.89 gram robot has achieved an 85-second, stable untethered hovering flight.

The CML team used the Objet Connex500, a multi-material extruding printer, because they needed to print very fine features for the wings. The 3D printer also enabled a wing-design cycle of a matter of minutes.

3D printing gives roboticists the ability to recreate biological structures, that would be just too costly and time-consuming to build using alternative techniques, says Lipson.

Pluses of 3D printers for robot design and manufacture outweigh negatives

?In particular, multi-material 3D printers that allow you to combine different materials with different mechanical properties enable you to reach new kinds of complexities, of the sort that you see in biological systems that are totally unobtainable using conventional manufacturing techniques,? says Lipson.

Despite rather expensive materials (specifically cartridge replacements) and some strength limitations in the finished product, for Lipson the benefits of 3D printers for robot design and manufacture ?far outweigh? the negatives.

Lipson, who made the front page of the New York Times in 2000 with a self-replicating, 3D-printed robot, contends that 3D printers are ?absolutely? the future of low-cost robot production.

?I have no doubt about that. Especially in the area of robotics, which involves relatively complex machines that benefit from the complexity 3D printers bring. And use of 3D printers will also grow if, as expected, robotics systems move into more custom-adapted shapes for various tasks, where mass production is not necessarily feasible.?

In the meantime, expect a mix of 3D printed and conventional manufacturing techniques to dominate, says Lispon, with a ?growing portion? of robot manufacturing switching to more agile, flexible manufacturing technologies over time because of their flexibility.

When dealing with the complex geometry of robots like the Ornithopter, there is no better production process available than 3D printing, says Ido Eylon, applications engineer at Objet.

Objet’s 3D printers range in price from approximately USD$20K to USD$250K, depending on the size of the build tray (Objet’s largest printer comes with a 500mm build tray) and the number of materials available to print with, says Eylon.

While admitting that 3D printers don’t generally offer the same degree of accuracy as traditional manufacturing methods, and the range of materials available to print with are limited, Eylon expects improvements in both areas over the coming years ?alongside reductions in prices for consumables.

As for using 3D printers in large-scale robot manufacturing, Eylon believes that ?we’re not exactly there yet,? but that ?at least some? robotic technologies will be suited to large-scale production in the future as the technology improves and costs fall.

RepRap originally conceived of as a self-replicating machine

The RepRap –an open-source 3D printer developed by Adrian Bowyer, a senior lecturer in the Department of Mechanical Engineering at the University of Bath– is an interesting case for roboticists to consider, since it almost certainly qualifies as a robot in its own right.

The RepRap was originally conceived of as a self-replicating machine ?one that can print copies of its own parts (minus the electrical components).

While its use amongst roboticists is largely confined to private individuals rather than large robotics companies, the RepRap can print with a wide range of materials (including ceramics) and with a typical resolution of about 0.1 mm, making it perfectly suitable for roboticists, says Bowyer.

?RepRap will produce prints as good as quality as, for example, something like a commercial FDM [extruder-type] printer of two or three years ago, and indeed better in some cases,? says Bowyer.

The Thingiverse website –where 3D printing enthusiasts share their designs– shows a wide range of robot designs aimed at open source 3D printers ?including grippers and the bipedal ‘Prodos’ bot.

The RepRap offers a competitive edge over proprietary models, Bowyer claims, because it’s free from the materials constrictions imposed by proprietary 3D printing companies, and the costs (self-assembly RepRap kits are available for around USD$1,000), are dramatically lower than those of their proprietary counterparts.

So, are 3D printers ready to revolutionize robotics manufacturing processes?

We can certainly expect the adoption of 3D printing technologies amongst roboticists to grow as more and more designers turn towards 3D printers for all their rapid prototyping needs.

But when it comes to manufacturing, the consensus amongst experts is ‘not yet’ –and almost certainly not in 2012. Nevertheless, robotics companies keen to gain the edge on their competitors would do well to watch the 3D-printing space very closely over the coming years ?particularly, looking out for 3D printers capable of delivering printed electronics at a reasonable cost.