Building a 21st Century Industrial Base via Open Source Technology

 

Forrest Higgs, PhD

Pacific Grove , California

 

 

How did it all begin?

 

 

...the only tools a dwarf needed were his axe and some means of making fire. That'd eventually get him a forge, and with that he could make simple tools, and with those he could make complex tools, and with complex tools a dwarf could make more or less anything. - The Truth , Terry Pratchett

Mankind did it rather like that. Our technological heritage developed over the ages by fits and starts until today we have a vast array of techniques and tools for crafting the things that make our civilization possible.

A whole literature of speculative fantasy literature centers around the notion of a modern man thrown across time and finding himself perforce having to reinvent the modern age by sheer dint of hard work.. Mark Twain's Connecticut Yankee in King Arthur's Court started the ball rolling over a hundred years ago.

 

 

The question came up again in a different sort of speculation in NASA on several occasions in the past forty years. The premise is that if mankind is to set up self-supporting off-planet colonies it is not going to be economic to ship more than the bare minimum of technological "vitamins", if you will, along with the colonists to allow them to build up a complete technological base.

 

 

 

 

Although the notion of self-replicating machines and technologies had been around for a long time the NASA studies looked at it in terms of purpose designing a whole self-replicating manufacturing technology with an eye towards keeping the starter kit as small as possible, interplanetary freight charges being what they are. The important thing that NASA brought to general awareness, however, was the fact that an overtly self-replicating manufacturing system could grow at an exponential rate of

~2^ n

 Where:

n - replication time

Some thirty years ago we began to see 3D printers, that is to say, printers that could print things instead of documents. At the onset they were fabulously expensive machines both to buy and to operate. As you might imagine you saw them first in the defense and aerospace industries. Prices have dropped over the years till today when you can set one up for about $30,000 and produce usable prototypes for about $30/cubic inch.

 

Two years ago, Dr Adrian Bowyer of the University of Bath in the United Kingdom undertook an endeavour that has been termed the RepRap Project, RepRap being a contraction of Rapid Replicating. After a period of milling around in November of 2006 Adrian developed what he has called the Mk 2 Fused Deposition Modelling (FDM) Extruder. Fused deposition modeling is a 3D printing technique originally developed by S. Scott Crump in the late 1980's. Crump's patents formed the intellectual property basis of the 3D printer company Stratasys, reputed to be the largest manufacturer of 3D printers in the industry.

 

Adrian 's Mk 2 extruder, seen here, represented a major breakthrough in the state of the art for FDM.

 

 

Crump's FDM extruders only work in a heated enclosure. This meant that the positioning system needed to maneuver the extruder had to operate in elevated temperatures. As you might imagine, that made it expensive. It also meant that Stratasys 3D printers used a lot of electricity to maintain those elevated temperatures. You see, "rapid prototyping", a term regularly used to describe 3D printing is a misleadingly relative term. FDM 3D printers are "rapid" when compared to programming a CNC machine tool and producing a prototype part in the usual way. They can take days or weeks to make a large, complicated part.

 

Bowyer's Mk 2 operates at room temperature. As you can see, it made mostly of plastic. Ironically, it was prototyped on a Stratasys 3D printer. This extruder draws about 15 watts of electrical power.

 

Stratasys' patents on the FDM process ran out last month. Dr. Bowyer could have patented the Mk 2 and used it as the basis for a new 3D printer company. Instead, he made the Mk 2 open source technology free for any use, commercial or personal, whatsoever. I begged my way onto his development team shortly after the Mk 2 was published on the RepRap Project website. Several of us began to develop prototype positioning systems for the Mk 2. Easily the most successful one is Vik Olivier's Zaphod.

 

 

 

 

Zaphod went hot in July of last year and produced its first replacement part in September at the Paraflows 2006 conference in Vienna .

 

 

 

In the past month Vik has started replicating and testing the much more critical parts kit of the Mk 2 FDM extruder.

 

 

 

 

Okay, that brings you up to speed on the technology. I showed you Vik Olivier's replicator largely because we have the most operating experience with it. It's time, however, to go back to basics.

 

For that, let's look at my prototype 3D replicator, Tommelise (Thumbelina).

 

 

As you see I made it out of milled poplar, a cheap quasi-hardwood that can be had quite inexpensively at virtually any sizeable home improvement store. I made mine with simple hand tools in my small seaside flat on the Monterrey Peninsula quite some distance up the coast from here. I designed and built it in my spare time over about three months last Autumn. The only bit that wasn't available at my home improvement store or on the internet was the Mk 2 FDM extruder parts kit. Dr. Bowyer was kind enough to supply me with that.

 

Mind, even that wasn't necessary given a suitably determined builder. Jim Wilkins made up a very serviceable Mk 2 FDM extruder parts kit out of wood and fiberboard last year.

 

To understand the point of doing this instead of just waiting for somebody to print you up a complete set of parts you have to look at the technology diffusion expression that I gave you earlier.

~2^ n

 

Where:

n - replication time

Tommelise has the equivalent in poplar of about 5 lbs of plastic. It can, running flat out produce about 8 lbs of extruded plastic parts/month. That means that it will take 19 days for Tommelise to reproduce itself. To suspect that someone would build a 3D replicator and then be selfless enough to run it full-time for years making kits for others is just silly. Suppose you could talk builders of such replicators to commit half-time on their machines to replicating parts. That would mean that you could make a full parts kit for Tommelise in 38 days or just at 10 per year. Using our diffusion expression that would mean that at the end of a year, if all Tommelise builders did that you coule have.

 

2^10 or 1024 Tommelises

 

at the end of two years you'd have a shade over 1 million. That's not a bad diffusion rate, but we can do a lot better than that.

 

With a reasonably equipped woodworking shop of the sort that exists in any high school and many, many garages you can rip out the parts for half a dozen Tommelises on a Saturday morning and assemble them in a day or so. The Mk 2 parts kit for Tommelise weighs about two ounces. Tommelise can print such a kit in 12 hours or two kits in a day. Suppose we only printed a kit out every two days

 

Looking again at out diffusion expression.

 

2^182 or 6 x 10^ 54 Tommelises

 

which is just a totally ridiculous number. Suffice it to say Tommelise as a emerging technology could diffuse across the world at a horrifying rate limited only by the availability of the parts necessary to make it.

 

Okay, so you can build a self-replicating 3D printer. Once you've done that and got tired of making parts kits for your friends and relatives, what can you make with it? The answer is . just about anything . Here's my personal favorite that was made in a Stratasys 3D printer that Tommelise ought to be able to duplicate.

 

 

Before you dismiss 3D printing as purely a plastics-bound technology keep in mind that it can also extrude clay into moulds for metals casting and extrude high technology ceramic slurries directly into precision parts for things like mini-gas turbines.

 

This all sounds beyond what an individual could hope to do till you realise that Stanford fired their turbine impellers in what was nothing more than a hacked home microwave oven that can be had at Costco for about $75.

 

Everyday individuals can produce a very large portion of the consumer items that they are presently buying out of China . Creative individuals can design new and/or better things right in their bedrooms.

 

Using a term mentioned earlier, what are the "vitamins" needed to build a Tommelise. Here's a look.

 

 

 

Besides poplar you need a few lengths of 3/8-16 threaded rod and 1/4 smooth rod than can be had from your hardware store. You also need a handful of electronics parts that can be had from any of several internet suppliers. Finally, you need several $6 plastic gearmotors of the sort that you can find on radio controlled cars.

Once you've put your Tommelise together you'll need feedstock. Here's what you need for now.

Basically, that's plastic welding rod which you can order in 5 lb lots for $4-5/lb, spackling compound to for support structures is readily available from any hardware store. While you're at your hardware store you'll want to get some plumber's or electrician's solder at about $7-8/lb. Did I mention that we expect for 3D printers to be printing parts with integral circuitry before the end of the year? That's not only wiring but also printed circuit boards. Finally, you'll have to replace the motors in your gearmotors from time to time. My supplier recently upgraded the motor he was using. The old one lasted about 3 months under harsh loads before the bearings on the motor, not the gearbox, wore out. He is confident that the new motors will last much longer than that. As it is, the replacements cost less than $2.

Tommelise is extremely energy efficient. It uses less electricity than two high efficiency light bulbs of the sort that the California legislature is trying to make mandatory here.

Tommelise looks like a cross between a joke and a toy. It is neither. It is a spin off of the RepRap project that I mentioned earlier. RepRap will be releasing their first 3D replicator next year. Here is, more or less, what it supposed to look like.

Darwin , is a very well-designed 3D printer. It promises to be able to print much more quickly than Tommelise, though perhaps a little less accurately. Because of this it will be able to make a full parts kit of itself possibly twice to three times as fast as Tommelise. It makes extensive use of steel rods instead of wood for structure.

The bad news is that you can't make one from scratch out of other materials. You have to have a full parts kit. It also uses stepper motors instead of gear motors, which means that it will be using two to three times as much electricity as Tommelise.

A full parts set for Darwin is currently estimated to cost around $400. Tommelise's parts set will cost less than half that much.

Again, it is easy to see these 3D printers as toys. After all they can print only about 5-10 lbs of plastic a month. What sort of industrial tool is that.

Consider. They print at virtually the same rate as commercial machines costing $30,000 and up.

My attitude towards the production rate issue is that if you want to double output from 8 to 16 lbs of plastic product per month, simply make another Tommelise, don't bother buying into more expensive technology to make it go faster

Let me give you a quick case study. I was first attracted to the RepRap project and the whole notion of 3D printing because I wanted and still want to design telepresence robots. Telepresence is a way of working at a long distance, with one's actions translated and transmitted over the interned and sensory feedback sent back from the 'bot. The image I have for such a 'bot has always been the loveable Johnny 5.

Over five years ago a surgeon in New York City did a gall bladder operation in France using telepresence. With a general purpose telepresence 'bot and operator harness you could fix a broken aircraft engine in the Congo or pick crops in California depending on your skill set and get paid via PayPal . all without leaving your home.

The problem for me has been the development costs. I know what's needed and how to do most things, but I am capital shy and from experience am averse to approaching venture capitalists until I have something working. Before I ran across the RepRap project I estimated that a cut-rate design programme for such a 'bot and user harness could easily set me back $50-60,000 and that assumed that I didn't make a lot of mistakes. I have a good track record for making lots of design mistakes.

You can design such a 'bot in a variety of ways and they all come out costing about the same. You can order parts from catalogs, you can set up a machine shop and make most of your parts, or you can buy a commercial 3D printer and design and make them that way.

The RepRap project, which led to Tommelise has given me a fourth way. I've estimated that with a working Tommelise I can do the design process for around $2-4,000. That's over a magnitude cheaper than conventional approaches to design. I can finance that out of my own pocket.

Further, when I have a workable telepresence 'bot I can go into limited production at the high end of the market by simply having my Tommelise replicate itself a few times. I can keep about 2-3 Tommelises in my flat. Once I want more to make more than 24 lbs of parts/month I'll have to wander down the street to my neighborhood storage facility and rent a few hundred square feet of space and scale up my production capacity, again out of sales revenue. Eventually, I can outsource production as sales justify.

What Tommelise, Zaphod and Darwin do is make the transition from idea to production virtually seamless and very low cost.

When I was a child, I spent a lot of time in our garage workshop trying to turn my ideas into hardware. I got a lot of things made, not wonderful things in most cases, but things nonetheless. The point was I had somewhere to make things and the tools to try. A huge percentage of kids today grow up with neither tools or garage. If they're lucky they don't have to share a bedroom in their parents' or single mom's apartment.

While these kids don't have tools or garages they virtually all have PC's. Tommelise is just another printer. Think of what could happen if kids, never mind adults, were able to design and make things and share the know-how over the web.