How DARPA Plans to Reinvent U.S. Manufacturing
DoD research wing wants to make everything from tanks to bombers in "fabs."
Christopher Mims 08/25/2010
- 15 Comments
 |
What if winning the Army's contest to design a new armored fighting vehicle was as simple as uploading your CAD files and other specifications to the U.S. Military, which would then have the capability to build your proposed vehicle - and all the other competing designs - in a generic fabrication facility?
If that sounds crazy, it's because everything the Defense Advanced Research Projects Agency proposes is deliberately out-there, from powered exoskeletons and synthetic blood, to robot insects and a flying submarine.
And DARPA has the budget to make these things happen, if they're even possible on a short enough time horizon: Americans are 4% of the world's population and half of its military spending, and DARPA alone has a research budget of $3.2 billion.
What DARPA is proposing is in its new iFAB program, for which it will soon ask for requests for proposals from private industry, is a "foundry-style manufacturing capability." By which they mean microchip foundries - the generic, build-any-chip-for-any-designer factories that churn out microchips for every application you can imagine, and which are the dominant mode of manufacture for most of the silicon in use today. (A handful of companies, like Intel, are rich enough to stick to the old way of doing things, and still own their own chip fabs.)
Here's how DARPA describes the program:
The specific goals of the iFAB program are to rapidly design and configure manufacturing capabilities to support the fabrication of a wide array of infantry fighting vehicle models and variants. Parallel efforts titled vehicleforge.mil and Fast Adaptable Next-Generation Ground Combat Vehicle (FANG) seek to develop the infrastructure for and conduct a series of design challenges (termed Adaptive Make Challenges) intended to precipitate open source design for a prototype of the Army's Ground Combat Vehicle (GCV).
 |
You read that right - the Army wants to open source the construction of its next-generation armored combat vehicle, the one that will replace its current workhorse, the Bradley Fighting Vehicle.
[Are you enjoying this post so far? If you are, help me expand into an hour-long talk for this year's SXSW tech conference in Austin. Vote for it as a panel here.]
If this sounds at all familiar, it's because companies like Local Motors are already trying to make this happen for everyday vehicles - custom design and custom manufacturing all made possible through what used to be called rapid prototyping and is now just making stuff in a big hurry.
The difference is that DARPA doesn't want to end up with just a bunch of kit cars - or in this case, kit tanks. DARPA wants to literally reinvent manufacturing - not just so they can build new vehicles more easily, but because they have a not-so-secret ambition to revive America's manufacturing base.
The iFAB end vision--to be developed in the second phase of the program which will be solicited under a separate BAA at the conclusion of the present effort--is that of a facility which can fabricate and assemble the winning FANG designs, verified and supplied in a comprehensive metalanguage representation with META/META-II tools.
Not everyone agrees that this is kind of manufacturing is a realistic goal - DARPA has a history of bringing on science fiction authors and futurists to help it brainstorm new ideas, and it's possible they were a little too high on a particular article from Wired when they wrote this document. But that's the point of DARPA - they fund things that no one else would, and eventually, that technology trickles down to the civilian sector.
 |
[ If you want to hear more about all the crazy science that DARPA is funding these days, vote to make it a panel talk at this year's SXSW tech conference in Austin. (Which, as you may know, is desperately in need of more presentations that are about actual technology, rather than just thinly-veiled promotions for the employer of whatever developers are sitting on the panel.) ]
Image of Bradley Fighting Vehicle courtesy Fort Knox Public Affairs Office
On Twitter
Become a Fan on Facebook
Subscribe to the Feed
DennisBuller
118 Comments
Manufacturing Issues?
I love DARPA, but I have a hard time seeing this working out.
The problems of US manufacturing is about labor cost and diffusion of capital and knowledge.
We train Chinese engineers here at our universities who then go back to China (diffusion of knowledge).
They then use the specialized training they received from us, and with the help of government money start businesses that compete with us. Then they ignore our patent rights and sell to the rest of the world, many times just copying our products.
So we lose on special knowledge and skills.
Their labor is much cheaper. So we cannot compete on price.
On top of this, there are many US companies that are moving their production facilities over there. Firing here, hiring there.
The replacement for the Bradly is going to cost a baggilion dollars a piece.
Why? Because there are few manufacturing companies that are left here that can do the job. The number of engineers we put out decreases every year. And the few companies that are left have the politicians running scared. They have become "to big to fail".
No matter how many times they screw up and do not deliver a decent product.....
You cannot "outsource" the design of something as complex as this. It is a not a software app. It requires hundreds of people with special skills interacting on a daily basis.
Does anyone remember the process of creation of the Bradly? It was so bad and such a boondoggle, they made a movie about it....
http://en.wikipedia.org/wiki/The_Pentagon_Wars
Having a "Fab" making the machine will not cut cost. Construction is the cheap part. It is the design that is expensive.....
-Dennis
www.PrometheusGoneWild.com
Reply
hellofu
9 Comments
Re: Manufacturing Issues?
yes and no changing from a standard production model wouldn't save money at first. you would end up making a more effienct system by being able to remove the majority of dedicated factories to a few high production general factories with a few dedicated factories filling the voids as needed.
this would cost more to start as you would have to make new factories but since you are starting from scrach you can redesign a factory so its fits a plug and play design like a lego set. this redesign could allow you to lower your work force yet produce more. this would most likely be done by adding more robots to the factories.
the change to Fabs would allow for a faster turn around to new products by companies allowing more work to pass thought the system making more products per man/robot hours and more money in ever ones pocket. this system has a better flow since it wouldn't need to reliant on just one assembly line to make a product as in a normal factory. this i Beneficial since if one key machine breaks it would stop production.
the solution to the problem with china would be to force a confrontation ending with non nuclear war. this would solve the problem of them stealing are trade secrets as no company would be able to directly buy or make products in china. eliminating a key economic base for china causing a slow if not a stall in there economy. it would end there ability to send chinese to the US for education. although a war would be costly it may be beneficial to the US economy in the long run as we would have to find new places to produce are products and most likely make a decent dent in are out sourcing.
Reply
gabrielg01
450 Comments
Re: Manufacturing Issues? - Chinese theft issue: shut China out.
Since the communist-nationalist Chinese are ungrateful for all the aid the US has given to them, we do have to change our stance toward them.
Just think of the monetized and social value of the higher education we have given to them. What is the economic value of educating hundreds of thousands of Chinese students? And how are they reciprocating?
Basically, they are taking advantage of our openness. This will have to change, otherwise they will steal everything.
The solution is to punish all our companies who do business in China (gasp!), and also ask for hefty tuition fees from all Chinese citizens who study in the US. No more free rides for those who do not abide by our value system.
Otherwise they will keep stealing everything "business as usual", no matter what DARPA comes up with.
Reply
cripdyke
52 Comments
Re: Manufacturing Issues?
In re: cost...
It would, in fact, save money. We know that because it works in the chip industry. Where the savings come from is primarily because the manufacturers can run at capacity all the time. There are many products that are manufactured for a relatively short time of popularity, or are purchased seasonally (think pet rocks for the first and lawnmowers for the second). A manufacturing plant that doesn't have to close down for a retrofit to switch from lawnmowers to snowblowers gains 2 weeks of productivity per year, and can instantly charge 4% less for its products.
But it gets better - there's no transition from one model year to another. Low-volume products can be produced on high-volume efficient manufacturing lines. Instead of a small, inefficient plant making dice 7 days a week, we can have a high-production, highly automated plant making dice for 1 hour per week, iPods for 3 days a week, frisbees for 2 hours, CDs for 3 days, and the last 3 hours they can make packaging for all the different products they made that week.
Packaging made at the same plant where the product is made saves shipping. When one business goes bust, there's no long period of inactivity while the factory is looking for a buyer and then being retooled for the next product. One customer doesn't need your manufacturing? Another customer will have an idea of something to manufacturer any day now, and with no capital purchase or long period of retooling/ramp-up, innovators don't need 7 million in start up funds. Just make 1000 units for $90 apiece. If they sell, then the next order can be bigger for $80 apiece and the savings can be passed on. Right now just the loans, capital & bargaining for a factory take a long time - which reduces overall economic efficiency though not manufacturing efficiency - and also idle any proposed plant and create a need to succeed big right away. There's no time for a product to slowly catch on & build word of mouth. A product whose advantages don't fit into a sound bite, if it is capable of being manufactured in an iFAB, can succeed without being made by 3M or Bose or Mitsubishi.
The constant stream of products to manufacture without retooling creates manufacturing efficiency. The rapid transition from old to new products creates economic growth efficiency as well as seasonal efficiency. The dial-a-yield manufacturing creates capital efficiency, as a new product doesn't need as much initial investment.
Don't hold your nose at this - it would be a major advantage to the US economy and its manufacturing sector.
However, this does not mean that China doesn't steal intellectual property willy-nilly. Many of the problems you state are real problems.
Just don't think that they mean that this movement toward iFAB-like production won't help the economy greatly. Economists talk about "capital efficiency." Its a difficult, complex concept, but the jist is this: when manufacturing slows (or the economy in general) the least efficient plants are taken off-line first, followed by more efficient plants if a recession gets worse, etc. The money, or capital, spent to create those plants is sitting idle. You can't sell the plant because the economy is gong to turn around and you'll want it then. But neither is your money being put to use making more money.
One of the worst parts of capital efficiency is that the plants taken off-line are the least efficient *by definition* not because they are *inherently* less efficient. What the definition really means is that the plants taken off-line are the ones that make the least money. This makes sense from an economic point of view, but not from a product-output point of view. What if people start buying fewer gas cans because more & more yard products are electric? A gas can plant might be very efficient at producing things, but still need to be shut down because it's not making money. It would be really nice if the plant could instead just shift from 7-days of gas cans to 6 days of gas cans and 1 day of battery-casings. The *product* was economically inefficient but not the plant itself. The plant itself is only inefficient in the sense that it is tied to a product.
When we decouple plants from individual products, we decouple manufacturing efficiency from economic efficiency.
If you know what that means, you know that that is an incredibly big deal.
Reply
DennisBuller
118 Comments
Re: War with China? (Response to Hellofu)
Good lord man, what have you been smoking?
I am annoyed the China is smart, and has taken notes from the Japanese on how to play us like a fiddle.
I do not wish them harm, and a war between us would be catastrophic and just plain stupid.
I just want a better deal and for the US to stop playing the fool.
Changing our tax code like the fellow below suggest would be a good start. Taking our patent issues to the world court would be another good step.
As for the fab saving money and being a great step in manufacturing; It may be. And after we have spent billions getting it right, other countries will steal the plans, or just hire the designers, and build their own.
At a much lower cost, since at that point all the issues will have been worked out....
Or better yet, the companies who get all this Federal Money to build the fabs will just start making them overseas since it is cheaper over there and will increase the profit margin.
And we will be still be dealing with the social structural problems with our manufacturing, just like we are now.
Not to be defeatist, I just think you cannot stop the ship from sinking by installing a better engine....
Don't take my word on it. What do I know?
But how about the CEO of Intel?
http://news.cnet.com/8301-13578_3-20014563-38.html
The comments on this article are funny; they break down into angry partisan talking points. They all have good points. But at the end of the day, will the CEO build a factory here or in India?
It will be built in India. And I cannot get a job there....
-Dennis
www.PrometheusGoneWild.com
Reply
tcm
1 Comment
Re: Manufacturing Issues?
I think that in time the Chinese will go the way of Japan, Korea and other states that have been producing for the USA - they will embrace capitalism, people will demand fair wages for a fair days pay and that will bring about change. It is ironic to me that the capitalists that have been chasing cheap labour around the world have caused this problem and denied their fellow citizens the right to work but by doing so have increased the standard of living for so many other countries.
Reply
Brian H
60 Comments
Re: Manufacturing Issues?
It's already happening to some extent. The labor component of manufacturing in China, and the clout of the skilled class of workers, is increasing fast. The "one child" policy has something to do with this; demographics are starting to bite.
Reply