Essay on putting printers in perspective
the future of printers
This is from a post in the forum, moved to the wiki to make it easier to find.
Some ideas and thinking about truly practical printers (especially putting printers into the big picture).
The reprap project, how it might fit into the larger world, and the future of printers.
You often hear and read people saying a printer or say a fab lab can make "almost anything" and "at the cost of materials".
And there are some real issues with those statements: when you start asking yourself what, of the things around you you can make, it quickly becomes obvious that that's a bit of an exaggeration. And the materials themselves are the product of human labor, just like products so why do they get special status?
I think this all gets much more interesting when you try to put it all in perspective:
precision machine shops
Bob talks about precision machine shops as a sort of self reproducing printer. If we expand the definition of printer to any piece or collection of equipment - such as a fab lab with the associated operators - that can produce usable real world physical goods then things get a lot more interesting pretty fast.
Complete Existing Printer
Humanity already has a self reproducing printer that can make literally anything in civilization, which consists of all manufacturing equipment on the planet plus all the human operators. Let's call it the Complete Existing Printer for now. Economics is pretty complicated, but it might be fair to say that this is the only manifestation that we already fully know how to do which can make everything in civilization *at the current prices we see on the market*. Although there is no doubt that a better printer is possible, which can make a wider variety of stuff more cheaply. Particularly because we have things like planned obsolescence and other features of the equipment, organization and end products which are inserted by counterproductive people on purpose and would be easy to do away with, providing large gains.
The variety and sophistication of the end products that the CEP can produce includes all materials, everything form Boeing 727s to spoons and semiconductor components, food etc. In a recent ted talk Tim Harford points out there are supposedly 10 billion distinct, different products available on the market for sale in New York City, for instance, so presumably there are many more elsewhere.
But what is the size of the printer? And how large and expensive is the share of it which is used by a single person in the developed world? I mean it can only be so expensive when you get right down to it because the total of society's infrastructure is owned by people (shares in corporations often), and the average person is not that wealthy, and the material goods production infractructure is a small fraction of that.
Small Almost Complete Reproducing Printer
Then, if you look at the smallest (in terms of capital cost) fraction of the system which - in isolation from the rest of the world - could produce say the 95 percentile or say top 5000 most economically important products at a price that is no greater than 20% of the market price for instance, and is still able to make itself, I bet that would be far smaller in terms of the amount of equipment needed because you do away with the million different types of merlot and so on, let's call it the Smallest Practical Almost Complete Reproducing Printer (SPACRP). Let's suppose we stick only with the type of technology, and costs measured as the current market price of the stuff, that is already being used. (Just a side note though that it would probably be really hard to predict what monetary prices of the produced equipment would be, so maybe it would be better to measure cost in terms of personel-hours and equipment time or something)
First of all there were many redundant components in the Complete Existing Printer. There are a great number of precision machine shops in the world, but the SPACP only needs one. Secondly, there are many production methods like injection molding that only exist to reduce the cost - for example pretty much anything you can make by injection molding you can machine directly, so some of those production methods would not be needed.
Thirdly, removing the 5 percentile least important products would probably greatly reduce the variety of stuff it needs to make, simplifying it. Would be extremely interesting to know what this printer would consist of.
ecology of equipment
The components of either the CEP or the SPACP could be arranged herarchically in terms of what you would use to make and maintain what - a CNC mill might be used to make a telescope but not vice versa. In many cases though the relationships end up coming full circle, you can make something with a mill which is part of the equipment you would normally use to make or maintain some part of the mill. Like a machine to make ball bearings, or a screwdriver to help assemble and maintain the mill.
Now imagine you take this further with a chart that showed all the relationships and dependencies between the different peices of production equipment - the ecology of equipment basically - remember it includes "land" as a sort of peice of equipment needed to produce food, chemical engineering plants needed to produce the raw materials like steel or plastic, etc. So it's not a tree, it's a very complicated chart although you might arrange the elements on the chart so the peices with the most stuff depending on them are at the bottom, like prey at the bottom of the food chain. And because it would end up including all the important stuff like housing and food and computers this SPACP could probably be the material basis of a pretty good economy.
Once you knew what it was there are probably a ton of places to step in and optimize it for what you want, especially if you are permitted to increase the net cost of producing some products by eliminating more of the cost-reducing-only production methods, or by tinkering with the design of the produced products and production equipment (which are themselves also products because it can make itself) to make them easier to produce with a smaller set of equipment and relatively few people and/or people who have less training.
In many cases you have a chemical plant or whatever and you might want to break it down into pumps, tubes, etc. (but with a border around those parts still) and represent it on the chart like that, which would make it more visible what parts of, say a ball bearing production machine, requires a lathe to produce. It might make the chart have less distinct elements in it, too, if you did that for all equipment, so you can see the number of particular types of parts involved which might be nice. It might be interesting to break it down into the individual moving parts of all the equipment, or in terms of the regions of the equipment that consist of homogeneous materials - so a painted peice of steel would be broken down into the paint layer and the steel - or various other things.
It might be kind of cool to visualize it as if it were displayed on google earth as an actual implementation of it, with the aluminum mine in one place, the steel mill in another, the machine shop somewhere else. Then click go, and it starts producing things, transporting aluminum to the mill to be converted from billet into sheet, then to the machine shop and so on. Each time material is moved you could draw a trace between the origin and destination, which had a few numbers displayed next to it that describe the person hours, whatever, that have been put into the transport process and the object being moved, or the value added after the last step or something. Maybe a number floating next to the destinations could indicate their cost similarly.
nucleus of an almost complete reproducing printer
Also in many cases the production equipment will be jigs or assembly lines needed to assemble an airplane or whatever which are specific to particular products but do not produce other production equipment. You could try excluding those so you end up with a sort of nucleus which is not itself capable of being an economy, but which could produce one given enough time and resources and good plans to follow (which is where open source would really help).
Also, when I say end products, almost all parts of the printer are also probably end products that the printer can make even if you did not deliberately include the "can reproduce self" criteria, as they will tend to be highly economically important, and almost all end products are also production equipment in way - the operators could wear the clothing they produce for instance. There may be some end products that are luxuries like wine or something that are only used very slightly if at all in the production process I guess.
other possible ways of shrinking or simplifying or improving the CEP
(a generalization of FutureToolIdeas)
You could further whittle down like this the amount of equipment and training needed to make a new model of printer you could call the Optimized Bang For Capital Expenditure Buck Printer. You could also try to reduce or eliminate dependency on remote mines or people that may not be there in the future, or whatever you want to.
Or even further whittle it down until you get the Printer That Costs Less Than $70k And Also Gives The Best Bang For Those Dollars Printer. The best bang in your view anyway. Or you could try to make the printer that had the least elements in the chart that could produce all the same products in isolation from the rest of the world, while allowing for price increases. There are lots of interesting things you could do with it.
At this point we could put fab labs, and the reprap and other new types of equipment or methods, in perspective easily. We can look at the SPACP or the PTCLT7AAGTBBFTDP or whatever your desired printer is, and see what peices of equipment the new kid on the block can replace most or all of the functionality of, and what, if you subbed in the new equipment for the old, the effect on the cost(s) of the products produced will be.
Often, when you are talking about say a RepRap, you want to know what vitamins it requires, and what sort of objects it can produce, what sort of effect it might have on the world. This approach of looking at a giant chart - an ecology of production equipment represented on paper - makes it easier to see the answer to those questions and put the capabilities of the RepRap model in perspective with the rest of the world, and the current production methods and equipment our economies already tend to use.
And also what it can make that the SPACP cannot, but let's face it, what humanity really needs is not so much different stuff or new types of stuff, but the capacity to make the stuff we can already make much cheaper, and secondly to actually use that capacity effectively for the right things, rather than what happens when it is in the hands of the bankster type people.
Then you could consider the importance of different types of theoretical new types of printer based on how they could change your main printer, like the CEP. The capabilities could be thought of in terms different resolutions and with different accuracy (and it's really accuracy that we ultimately want, not precision) and materials, and printing speed. There is the printer that can print in only one material to 2 micron accuracy or so with several micron voxels, like a direct metal laser sintering machine, and we already have those but they cost too much right now. There is the printer that can print something with a small range of materials in any desired voxel. There is the printer which can put whatever material you would like in any voxel (volume pixel). Then there is the molecular assembler, which is like a super high resolution of the latter because it can make something with atoms placed wherever you want. And you could go even further to subatomic particles I guess.
There would be many variations in between and probably a lot of practical printers, certainly the early ones, will be somewhere in between. Suppose a 2 micron voxel 1 micron accurate 3 material printer could also control some of the metallurgical properties of the metal it was depositing for example, like temper some areas of a high carbon steel part. That is having limited control at an atomic level but with similar or likely lower resolution. Or it could control the properties which the boundary between different material voxels had - a strong bond or a very weak bond, which essentially the ability to exert limited control at a low level.
In practice, at first, there will likely be many different limits on the printing capabilities besides voxel size and material types - you may not be able to get as high a precision for where exactly the boundary between 2 different materials on the workpiece is as the precision for where the boundary between a single material and the air (or surrounding vacuum or powder bed) is. It may be possible to define the boundaries of your object to within say 3 microns, but the radius of curvature may be more limited - due to the inertia of the galvanometer mirrors for example. Or the rate at which the intensity of the laser can be modulated may produce limitations of it's own for another example.
Or maybe you could control the orientation of the molecules within the voxels, so you could produce a material like Dyneema from Ultra high molecular weight polyethylene, Dyneema is a material in which the UHMWPE molecules are mostly aligned lengthwise along a particular axis, making if far stronger along that axis. Or the orientation of fibers in a composite material maybe, but most composite materials have fibers bigger than 1 micron, so you could also print a composite material like glass filled nylon directly. Except printing say a glass directly adjacent to a plastic (after the plastic) might be problematic for the heat-sintering sort of approach because the melt pool temperature of the glass is too high for the plastic to stand without burning.