Self-replicating rapid prototyping machine builds on its own success

Paul Boughton

Rapid prototyping and 3D printing are becoming commonplace, but the development of a self-replicating machine is making it possible for anyone to have their own machine for around the same price as a high-specification mobile telephone. Alistair Rae reports.

There is something truly wonderful about a seed growing into a plant that produces its own seeds - from which more plants can be grown. In the world of science fiction, authors have written about self-replicating machines that reproduce and take over the world. But the concept of a self-replicating machine is now closer to reality than ever before, with one group of researchers, based in the UK's University of Bath, claiming that in May 2008 the first 'child' was produced from a 'parent' machine, and the first 'grandchild' produced by the 'child' machine shortly after the 'child' was assembled (Fig.1).

Dr Adrian Bowyer initiated the Reprap (replicating rapid-prototyper) project in 2004, with the aim of developing a rapid prototyping or three-dimensional (3D) printing machine that would be capable of manufacturing the majority of its own components. Builders of such machines would only need to source readily available and relatively low-cost components such as screws, bushes, lubricant, a power supply, stepper motors, and electronic chips and sensors. Human input would also be required to assemble the components, of course. An important aspect of the project was that the component designs and control programs would be released free of charge under the GNU general public licence.

At the time of writing, the Reprap1.0 Darwin machine is available in this format, with the components costing in the region of EUR400, which compares with around EUR19 000 for a conventional rapid prototyping machine. Of course, before anyone can start building a Reprap machine, they first have to source a set of components. To avoid this from becoming a problem, the Reprap community has established an etiquette in which each machine builder is asked to use their machine to make the parts for at least two more machines for other people at cost. In addition, a company called Bits From Bytes is selling a set of Darwin components that are laser cut from acrylic, as well as some of the standard components required to complete a machine and consumables for building components on the finished machine.

Unlike commercially manufactured products, in which there are costs associated with tooling, the critical components of a Reprap machine are built in a 'batch of one' and there are therefore no set-up costs (other than design time) associated with producing modified versions. It is therefore expected that the Reprap machine design will be refined very quickly, and old machines will, it is hoped, be capable of producing new parts to the latest designs so that they can be readily upgraded for very little cost. The intention is that the designs for these new components - as well as designs for other products manufactured on Reprap machines - will be made available free of charge to the Reprap community via the internet.

As well as anticipating an evolving design for the Darwin machine, the Reprap team is already considering Reprapv2.0, to be known as Mendel. This will be similar to the Darwin, but capable of printing conductive materials as well as polymers, thereby enabling embedded electrical circuits to be constructed. Other proposed enhancements include a fourth axis, in the form of a turntable, to simplify the creation of round parts, and Reprap components to replace some of the off-the-shelf components required for Darwin.

Machine design

Darwin is a 3D Cartesian robot with a print head that extrudes 0.5mm diameter CAPA (polycaprolactone), PLA (polylactic acid), HDPE (high-density polyethylene) or ABS (acrylonitrile butadiene styrene). The robot is constructed from rapid-prototyped (or Reprapped) parts and steel rods, and the print head is built mainly from rapid-prototyped (or Reprapped) parts (Fig.2). In operation, the print head is moved in the X and Y axes, with the machine bed moving in the Z axis.

Although the goal of the project is to create a machine that can replicate itself, the machine's usefulness extends much further. On the Reprap website there are examples of products and components that have already been manufactured on Reprap machines. These include a fly swat, water filter, a pair of children's shoes (Fig.3), a door handle, a coat hook and a bracket to mount an iPod on a car dashboard. This last example is a good illustration of how Reprap machines can be used. It is a relatively simple design that solves the problem of how to mount a particular model of music player to a particular model of car - without the compromises often associated with universal brackets that attempt to fit all music players to all cars. One final example of a Reprapped product worth highlighting is the Reprap shot glass. According to the Reprap website, it has become a tradition that the first item made on a new Reprap machine is the shot glass, so that the machine can be toasted with something that it has made.

So far the Reprap project has appealed mainly to individuals and research organisations. However, it should be remembered that the concept has the potential to be extremely powerful. Adrian Bowyer uses the example of an injection moulding machine that is capable of producing 10000 combs per hour. If a Reprap machine produces one comb and one copy of itself per day, within 18 days the exponential growth in production capacity means that the family of Reprap machines will be producing combs at a faster rate than the injection moulding machine. This may be an extreme example, but it illustrates how Reprap machines could transform manufacturing and the quality of life in developing regions of the world.

On the other hand, in the developed world Bowyer envisages Reprap machines being used to manufacture replacement parts for domestic appliances, with the user having downloaded the appropriate CAD files from the internet. People at home or at work could also download files for new products, with some files being available free and others available for a fee (though the ease with which music files can be shared suggests that a pay-per-download business model may not be the most robust).

With parts being so easy and cheap to manufacture, it is possible that users may manufacture products or components that are not long-lasting, or that are quickly superseded by improved versions. This could result in large numbers of broken or redundant parts creating a waste problem.

However, as Bowyer points out, a Reprap machine could also build many of the parts necessary to create a means of recycling Reprapped products and process waste.

In addition, such a recycling unit could process post-consumer waste from other sources - such as HDPE food containers - for use in a Reprap machine.

While this offers one source of low-cost, environmentally-friendly material for a Reprap machine, there is also another that is causing excitement in some quarters. Polymers such as polylactic acid can be made from biomass by bacterial fermentation.

Starting with a few tens of square metres of land, Bowyer suggests that a starch crop (maize, for example) could be grown, enabling Reprap users to make their own polymer in a fermenter made using Reprap machines. This method of material sourcing also has the advantage that it takes carbon dioxide out of the atmosphere and stores it in plastic products. At the end of each product's useful life, the material could be recycled, though polylactic acid is also biodegradable.

Reprap machines are not yet challenging conventional rapid prototyping, rapid manufacturing and 3D printing technologies, but the pace of development is likely to be fast, at least initially. It will also be interesting to see which types of user make the most of the potential, whether they are hobbyists, product designers or manufacturers in the developed or developing world.

Recent Issues