Next stop: Hardware piracy
Publication date: 12 December 2010Originally published 2010 in Atomic: Maximum Power Computing
Last modified 03-Dec-2011.
I last wrote about rapid prototyping five years ago. There've been some changes.
Rapid prototyping is what it sounds like - making things, faster than you could do it by hand. Usually to make prototypes of something that you then intend to manufacture in a normal factory.
(Technically, it's only rapid prototyping if it's done by "additive" sticking-together of some cohesive material, not by "subtractive" carving away of parts of a block, which is what conventional machine tools do. Even an un-computerised machine-shop can make certain kinds of prototype pretty darn quickly, but it probably won't be advertised as a rapid-prototyping shop.)
There's a rapidly expanding rapid-prototyping enthusiast community, who often aren't actually making prototypes. They're more interested in making small runs of final products, right there on the desktop. People have started calling this "fabbing", and there are now several "digital fabricator" - also known as "fabber" or just plain "fab" - projects that're suitable for hobbyists. (The biggest names in relatively-cheap DIY fabs, so far, are "RepRap", "Fab@Home" and "MakerBot", but there are several others.)
Early efforts in small fabbers were largely focussed on making parts for more small fabbers, but now plenty of fabricators are making other things.
A "fabricator" always used to be a person who fabricated things, just as a "computer" always used to be a person who computed. The meaning of "computer" slowly shifted as mechanical adding machines developed into electronic computers, and now the meaning of "fabricator" is starting to blur, too. Someone who hand-builds custom exhaust manifolds is still a fabricator, but if you're reading Hack a Day or the Make blog or, of course, that one whole issue of the Make magazine, and see a fabricator, it'll be a little machine. Which, usually, extrudes a thread of hot plastic to build up an object, some or all of which is likely to have a distinctive scribbly look.
There are also small fabs that print optically-hardened resin. Special mention must also go to the Evil Mad Scientists' CandyFab, which prints by fusing together granulated sugar with a heat gun. Conceptually, this is similar to "selective laser sintering", which does the same thing to a powder with more desirable physical properties than sugar.
Sintered-metal 3D prints are often rather fragile, but they can be dipped in another metal with a lower melting point to fill the tiny voids, and that can leave you with a very sturdy object, like several of Bathsheba Grossman's sculptures.
(You could probably make a home laser sinterer that fuses zinc powder, then dip the resultant zinc-sponge in molten tin for a moderately durable result. Since I have no intention whatsoever of trying to do this, I'm quite confident that it can't be very difficult.)
Commercial rapid fabricators already exist, some with capabilities approximately as amazing as the computerised milling machines that can turn a blank slab of alloy into an engine block. But professional fabbers are extremely expensive to buy and run. The big news is small, cheap fabs.
And big news it really, really is, in certain circles at least. There's an absolute mania for the things. I kept adding more and more and more examples to my source file for this article, until I finally had to stop, to actually get this thing written. I swear that for a while there, there were more 3D printing posts on maker blogs than Minecraft posts on game blogs.
(Fortunately, as I write this, everybody's gone nuts for Kinect hacking, briefly slowing the avalanche of fabbing news.)
To pick just a few arbitrary examples: Puzzle pieces. Tool holders that exactly fit your tools. Various glowy things. Numerous parts-storage systems. That pocket-sized branding iron you always wanted. Yet more iPhone accessories, because heaven knows the world doesn't contain nearly enough of those yet. Models of space-ships from your computer game (it's often quite simple to use the 3D data that renders an object on the screen to render it in the real world). Even ceramic pots.
It's gotten to the point where the Make blog, the other day, ran a 3D-fabrication holiday gift guide.
Atoms are not really "the new bits", if only because you can't duplicate atoms for free or distribute them with wires. But there are still definite similarities between fabbing atoms and manipulating data.
So a surprising amount of manufacturing capability may, soon, be sitting on ordinary consumers' desks.
This sounds ridiculous on the face of it, but it wouldn't be the first time something this peculiar has happened.
For a mundane example, look at desktop publishing. The very concept of designing and printing your own professional-looking documents at home was ludicrous, until suddenly consumer-market laser printers and graphical-interface home computers made it easy.
Home computers are actually a good example all by themselves. Old science-fiction is full of ray-guns and sentient robots and faster-than-light drives, but you won't find any personal computers. Because, in 1955, the very notion of a "personal computer" was about as plausible as a "personal ocean liner". It seemed pretty probable that people in 1985 would all have personal helicopters, but personal computers? Pull the other one!
Thinking about it this way, there are plenty of situations in which comparing home 3D printing with the old-fashioned alternatives would feel like comparing home document printing with putting the image on paper with a pen.
Spare parts, for instance.
You know how you can go to an abandonware site, or Good Old Games, and grab yourself a free or almost-free copy of some excellent old game?
Well, how about doing the same thing for parts for your old car, or old dishwasher, or old toys, or old pretty-much-anything-else?
You won't be able to make a new controller board for a modern computerised appliance, and things with moving parts or specialised materials may also be a problem (like, for instance, a new fan module for your fan-forced oven...). But home fabs can already make a variety of more prosaic items.
In both of the above videos, and in most other fabber projects, a fabber is obviously not the only possible way of solving the problem without buying a pre-made part. Even if you ignore non-optimal solutions like replacing a releasable clip with glue or double-sided tape, you could still carve a replacement out of wood by hand or with machines, or make something out of polycaprolactone by hand instead of by extruding the same plastic from a low-temperature fabber nozzle. And for applications not involving live wires, it's amazing what you can make with a bit of thought, some coat-hangers and a wire-bending jig.
But home fabs won't need to be much better than they are already for most plastic hardware-store fixtures and fittings to become fabbable. Fabbing is still not a winner for bulk production, but if you need a new plumbing fitting or shelf supports to fit a given spot perfectly, or some other item that'd have you stalking the aisles at the local Hardware Blimp Hanger with notes and a tape measure, then just whipping up the thing you want in some SketchUp-like interface and then clicking the "Fab!" button becomes immensely attractive.
There's also no law that says you have to fab every part of something. Future home fabs may be able to construct certain circuit boards, probably some kinds of battery, perhaps even transistors - but a lot of that stuff is going to have to be off-the-shelf components made in a normal factory and then incorporated into a fabbed whole.
This is actually the way normal manufacturers work already. Only if a product has very specialised requirements is it necessary for the manufacturer to create a whole new electric motor, bearing, microcontroller or what-have you to put in it.
If you want any sort of home fabber for less than about $US3000 ("serious" commercial fabricators cost a lot more), then you have to build your fabber at home. But we're getting past the stage where every cheap rapid prototyper was, itself, a prototype and constant work-in-progress. Now, we're starting to see off-the-shelf fabber kits like the MakerBot Thing-O-Matic, which you can pretty much just buy, build and use, like a loudspeaker or R/C-toy kit.
(Oh, and there's currently some money waiting for whoever manages to build a better RepRap.)
(Printer companies are also starting to show interest in making true pre-built off-the-shelf fabbers affordable by home users. HP's commitment to the idea has thus been limited to agreeing to rebadge someone else's product - but still.)
The sci-fi-story dream fabber takes a blueprint for a mobile phone, crossbow, turbocharger or original G1 Walther P-38 Megatron, chugs away for a few hours, and then delivers the toasty-warm, just-like-a-bought-one item into your hands.
Don't hold your breath for that to happen. You'd need cyberpunk nanotech to make it feasible, plus either a ridiculous number of different feedstock materials, or a tame alchemist.
But compared with the rate of progress in most industries, fabbers are developing at a lightning pace. If it looks slow, it's just because we're all spoiled by the outrageous rate of improvement of computer hardware.
I wouldn't be surprised to see home fabbers printing arbitrary Lego components some time in the next five years.
This would be more important than you might at first think. Current home fabbers can make things that're just barely compatible with Lego, but they don't yet have the precision to do it properly. Once you can make commercial-quality Lego, though, you can also make a long list of other quite serious engineering components. Home fabs are getting there.
Prices for kit-built desktop fabbers are already dipping below $1000. The bulk plastic most of them extrude isn't expensive, either. They're definitely more accessible to the home hacker - or, at least, easier for the home hacker to lift - than many standard machine tools.
The specialised Web sites are showing up now, too. MakerBot Industries' "Thingiverse" is your one-stop shop for free design swapping and collaboration, and "BatchPCB" is essentially Zazzle for circuit boards. 3D scanners are slowly getting better and cheaper, too. By the time it's unremarkable for hobbyists to have a 3D printer in the garage, they'll also be able to easily scan that broken piece of plastic that's stopping a gadget from working, and then fab an un-broken version, probably from a tougher polymer.
We may never actually end up with fabs as unremarkable household appliances. Look at printers for a comparison: There've been true photo-quality inkjet printers for many years, now, but countless users have given up on printers which, unless you use them frequently, need a ten-minute-and-twenty-five-dollar cleaning ritual whenever you want to print one darn photo.
The alternative to a-fabber-in-every-garage is the "output bureau" model. There are already services that'll take a 3D model and rapid-fabricate it in a variety of materials, just as you can get your digital photos printed at a kiosk or via an online service. I.materialise and Shapeways, for instance, both offer far better precision and far more materials than you can get out of a hobby fab.
(I also like the idea of a travelling factory, like a cross between nomadic tinkers and the circus.)
I think you're going to have to use a service like that if you want to print something in metal. If you want a metal object, you could print it in plastic and then use the plastic as a casting pattern, but casting is a whole separate field of endeavour.
ABS - the stuff Lego's made from - is probably the most popular plastic for extruder-type fabs, and it melts only slightly above the boiling point of water. There are low-melting-point alloys that a cheap desktop fabber could probably cope with, but those alloys all have such unexciting physical properties - poor strength, toxic constituent metals - that you'd almost always be better off with plastic. To extrude lead-free pewter, you'd probably need a 300°C nozzle.
It's not, however, out of the question that hobbyists will soon be producing even quite complex items, like circuit boards, in a pretty automated way. People are already, for instance, working on hobby-scale pick-and-place systems, like this one.
There's also not much distance between a fab and a laser cutter/engraver. The two concepts go together rather well. It's not hard to imagine a convertible combination machine that does both jobs.
(A lot of kit-fabs are actually already made from laser-cut materials. The black edges and plain faces of laser-cut MDF are as distinctive, and as common, as the scribbles of extruded-plastic fabber output.)
If small-scale manufacturing becomes really widespread, the effect could be very dramatic indeed. It could be one of those great reducers of technical friction, like the development of standardised fasteners.
I'm also, however, waiting for some great moral panic about fabbing to kick off. I don't think most fab-enthusiasts have much interest in making slingshots, crossbows and smoothbore automatic air-guns (not to mention their own unmanned aerial vehicles with which to deliver the poison gas they made with instructions from hackers on steroids and chemicals from mysterious sources), but you can bet some fab-enthusiasts are going to do it.
It probably won't be long before home fabs are able to make most of the parts for a gun. Blokes in Afghanistan have for many years been making arbitrary firearms from scratch with little more than hand tools. So fabbed zip guns can't be very far away, and Stories No Parent Should Miss on the subject may be rather closer.
(For this reason, it's possible that capable fabs and/or their feedstock materials will be legally restricted in some countries. John Scalzi's "The Android's Dream" was somewhat ahead of the curve on this subject.)
I don't really know where home fabbing's going, though. I suspect that while we're all distracted by people making Nerf guns and Stormtrooper armour, someone's going to startle us all with something like a $20 home-built DNA computer.
People are starting to make their own printers, too. This seems a bit silly, until you realise that "printing" can involve a lot more than just ink on paper. Printed circuits are a possibility, of course, but also e-paper, solar cells, cake icing, skin grafts, whole human organs...
Or maybe someone'll come up with a price breakthrough in 3D microfabrication. Commercial focused-laser microfabs can make mechanical parts below micron scale (for comparison, the micromirrors in a DLP projector are more than ten microns across). Soon, hobbyists could be making autonomous robots the size of sugar cubes.
One of the big reasons why it's hard to invent new things is that the inventor often finds themselves having to make new tools to make the device, or even having to make tools to make the tools, et cetera. And once an inventor finally gets past these hurdles, they're usually not particularly inclined to now go into the tool-making business to help other people over these hurdles.
Today, the Web lets people share how-to information with the world. In the future, widespread access to fabbers will extend this collaborative data-sharing into collaborative object-sharing, which should massively streamline the production of many kinds of esoteric equipment. The production and operation of fabbers themselves is, of course, a tools-to-make-the-tools situation, but every day more tips and tricks circulate in the fabbing community.
Whatever happens, one thing's for sure:
If you sell collectible action figures for a living, you should be nervous.
