Blog Menu

Machines that Build Themselves

January 5, 2008

Filed under: additive manufacturing,future,manufacturing — Terry Wohlers @ 09:28

Self-replicating machines have been a topic of futurists and science fiction writers. Nanotechnology shows some promise for nanoscale assembly, although practical applications of this may be many years into the future, if ever. A professor at the University of Bath in England launched an ambitious open source project a few years ago that aims to produce a macroscale self-replicating machine by additive fabrication (AF), although little evidence of actual self-replication has been demonstrated thus far.

Today, two companies offer machines that are beginning to build themselves. One year ago, EOS announced that laser sintering was used to produce 23 parts on its Formiga P 100 laser sintering system. Among the parts being produced are the filler hopper for the plastic powder, a switch cover, and pieces for a pyrometer. Last month at EuroMold 2007, Stratasys announced that fused deposition modeling (FDM) was used to manufacture 32 parts for its new FDM 900mc system. Some of the parts include the touch screen bezel, door latch filler, pull handles, status tower base, and cable strain relief bracket.

As the capabilities and materials for these machines improve, expect the number of parts that they build for themselves to increase. Will they ever be capable of producing themselves entirely? Maybe someday, but not until systems can process a very wide range of materials, including plastics, composites, and metals. Today’s machines can process plastics/composites or metals, but not both. For a long time into the future, standard parts, such as motors, gears, bearings, belts, wires, printed circuit boards, switches, fasteners, and sheet metal, will be purchased and assembled the way they have in the past.

1 Comment

  1. I wonder if you know about the RepRap project. Its aim is to design, through the open source process, a 3D printer that can replicate most of its own parts. Currently a large fraction of the parts can in fact be 3D printed, save some nuts and bolts and rods (60+ unique parts, little over a hundred of parts in total are printed). This fraction won’t increase much because it’s already met a point where fabbing the supporting rods is not desirable (total costs are very low, they are available anywhere), but it could be done nevertheless. Same goes for screws and such.
    What is still desirable to be able to replicate is the electronics. In fact, the standard (1.0) model of the RepRap, the ‘Darwin’ is capable of milling professional-looking and absolutely functional PCBs for the electronics

    Some initial successes are in fabricating the stepper motors to control the motion of the machine. The need for fabricating readily available stepper motors, that can even be scavenged from dead/outdated 2D printers, is not that pressing to fuel this avenue of development, but it certainly is possible that we will 3D print the basis for functional motors.

    RepRap can also function as a bootstrap machine to 3D print parts for other machines than itself. Many rapid prototyping machines, can produce parts for a RepRap (there are arcylic, steel, plywood and FDM versions of most parts). RepRap is more than just the machine, it is also the toolchain in software. Most of all, it is the helpful in very thriving, knowledgable multi-disciplinary community.

    For more information, check out

    Comment by ErikDeBruijn — October 21, 2008 @ 03:03