Open Source Ecology

A day or two ago I heard a piece on NPR about a guy named Marcin Jakubowski who is on a mission to design and build the Global Village Construction Set, “a modular, DIY, low-cost, high-performance platform that allows for the easy fabrication of the 50 different Industrial Machines that it takes to build a small, sustainable civilization with modern comforts.”   As described in a short TED talk, Jakubowski explains that he was a theoretical physics student, started farming when he felt that he lacked practical skills, and had bad experiences with old farm equipment that motivated him to design simple, practical open-source tools. This evening I checked out the website of his organization, opensourceecology.org,  which has a work-in-progress wiki-based compendium of concepts, specifications, designs, fabrication videos, and test images.  The group is big on modularity, hydraulics, and structural steel, and they’ve done a pretty impressive amount of development on a few of their concepts, including a skid-steer tractor and an automated press for making rammed-earth blocks. They are based at Factor e Farm in Missouri, which reminds me a lot of stuff I’ve read about the New Alchemy Institute.

OSE is an interesting concept, and they’ve come up with some cool stuff – they’ve turned their rudimentary skid-steer into a rudimentary crawler by welding up tracks out of ordinary chain and ordinary rebar.  They built functioning mobile dimension saw that runs on the same hydraulic power unit as their tractor.  And they have a pile of ambitious concepts that are partially sketched out in a sprawling wiki.  Clearly there are some capable and energetic folks involved, and they’ve managed to secure tools, a materials budget, and the ability to produce nice videos of the stuff they make.

It won’t come as a surprise that this whole concept has an inherent appeal for me, and there are so many directions to launch off in a discussion about it.  But quickly my natural skepticism kicks in, and asks, ‘in what applications and under what circumstances is this actually a workable model?’ And (cutting to the point) I very much doubt that open source microscale equipment is going to revolutionize the means of production, but do think that thoughtfully conceived and carefully tested and refined  open-source designs could actually become significant, in specialized applications where scale issues limit the massive advantages of modern manufacturing technology.

First off, I dismiss the idea that the open source ecology toolbox as currently conceived (even if every piece in the toolbox were perfected)  is a realistic general strategy for practical homesteading, eco-villaging,  third-world development, or navigating civilizational collapse.  I have some few years under my belt now as a spare-time casual homesteader (between Five Islands and Gorham) and the last thing on my list of things to do would be to build my own tractor (which seems to be the tool that OSE has spent the most time on, and which seems to be the most well-developed, excepting the earth-rammer).  Used tractors are readily available, at every point on the continuum between moribund and quite cheap to relatively recent, quite reliable, and nonetheless substantially discounted from the off-the lot price.      In Five Islands we have two small Kubota tractors, one of which was bought used 30 years ago and has been in regular service since with minimal repairs or maintenance. Having built a lot of prototypes of a lot of things, and being a respectable amateur machinist and fabricator, I can say with confidence that I could not build a homemade tractor of equivalent functionality, reliability, and longevity from general-purpose starting materials at a cost in time and materials-at-market-prices even close to what I would spend on a used diesel tractor – even if I had expertly crafted and tested open-source plans in hand. Even if I had no money to buy a used tractor and my time was free, how then would I buy the stack of steel and hydraulic components, engine etc. that I would need to get started? Not to mention the several thousands of dollars of equipment I would need to efficiently do the fabrication.  I would be better off bartering with a neighbor to borrow her tractor, or doing some work in the cash economy so as to buy or rent a tractor.  And we’re talking about someone who already knows how to design, cut, weld, machine, fabricate, assemble hydraulics, etc.  If I had to learn some or all of those things before I could get started, that makes the extreme DIY tractor concept even crazier.

The reality is that modern manufacturing is phenomenally efficient, and small farm tractors are a solved problem.  In that way they are a lot like bicycles.  Ever since I learned to TIG weld, I’ve always thought it would be cool to weld my own bicycle from tube. But I’ve never actually done it, basically because bikes are cheap and awesome: I paid $200 bucks for a used Miyata touring bike in college, it fits me fine, and I put thousands of miles on it with nearly zero maintenance.  It would probably take several attempts and a lot of scrap before I could make something that was even marginally better than an extremely satisfactory product that I can find at a very reasonable price in a yard sale. Making bikes for fun is cool, and good on anybody who wants to do it, but I can’t see why an open source bicycle should be high on any priority list – and similarly tractors.

I don’t see how the ‘solved problem’ aspect changes much  if you’re considering some sort of hypothetical eco-village.  Setting aside the problems of how to capitalize, launch, organize, and manage an eco-village, the same type of tradeoff is going to come up – do we put our limited resources into building custom equipment from scratch, or do we buy conventionally manufactured equipment (used or new)?  Here the cost of fabrication tools can be amortized over a greater number of devices that can realistically be built, and we can afford to have specialists who do nothing but fabricate equipment.  Still, I believe that the optimum solution would be to buy conventionally-manufactured equipment where it is available, build a decent shop for repairs and modifications, and focus available innovation/fabrication capabilities on specialized equipment not otherwise on offer for a reasonable price.

Besides growing up off the grid in the pre-PV era, I don’t really have any personal experience with developing-world technology, but the sense I get is that the challenges are very situation-specific, and a general-purpose industrial toolkit based on the availability of standard stock and components is no more likely to dominate in the developing world than in the context of our lives.  That brings me to the question of how OSE would fare in some type of hypothetical collapse scenario, and I would argue that things aren’t looking good for our heroes.  Leaving aside that most of the stuff they’ve built runs on gas or diesel, much of the value of the open-source paradigm as they’re laying it out is that it’s built from readily available standard stock and parts.  But in any kind of disruption severe enough to require a buildout of microscale production equipment, I would not give a lot for the probability of the ready delivery of mild steel structural tubing and standard hydraulic fittings. I totally understand the romantic appeal of building the world anew and afresh from double-strength square tube and 6011 rod, but I have to think that in any real scenario, things would tend more toward ‘Dude! There’s an old Buick in this garage – I bet we can get it running and build a doodlebug!” (Or, if your fantasy runs more toward steampunk, you may alternatively begin the sentence with “Hark!”) OSE hints at some of these issues by including a placeholder page for an induction furnace, but there are huge gaps in the story (e.g. “blast furnace” and “rolling mill”).

But, even if OSE isn’t going to immediately revolutionize the means of production, I think it could be incredibly useful.  And I think one key is to recognize that modern manufacturing does some things incredibly well, and it doesn’t make sense (at least under present circumstances) to go head to head with it.  But there are still a lot of niches where there’s plenty of room for improvement.  I think their rammed earth press is a great example of this.  I have heard of people using rammed earth bricks, but I have never seen them available for sale, and I have never seen a rammed-earth brick press available for sale at an ag supplier.  I don’t actually know how useful rammed earth is as a building material, but I’ll imagine that in specific applications it’s great.  Probably those applications are too rare or specific to have triggered conventional manufacturers to commercialize the pressing equipment, and that seems like a fantastic application for open source hardware plans: where a person of moderate skill can build something useful and truly state of the art.

I would argue that cider-making is another example where volume hasn’t justified manufacturing innovation, and we have an opportunity to make nearly state-of-the-art equipment (at least at the small-scale).  I think the stuff we’ve built is a pretty good start, and we’ve published plans and instructions for some parts (I would be happy to send the entire solid model to anybody who wanted to organize and open-source it).  In general, the less modern manufacturing has been brought to bear in standardizing and driving down the cost of a particular tool, the more opportunity I see for open-source hardware to really make a difference.   For instance, modern portable sawmills aren’t built much differently from the OSE fabrication paradigm – standard steel parts and mechanical components are cut and welded together to make clunky-but-robust machines that run on small engines.  In some cases a few custom castings or the like are probably used in places where it enables a bunch of functionality to be packed into a smaller space or lower cost, but surely a skilled fabricator working from good plans could build a product that’s about as good as a Woodmizer or Mobile Dimension saw. And depending on how efficient they were and how much they valued their time, they might be able to build it for less money than buying a new unit – given that they’re getting the design for free, and they don’t have the overhead, marketing, and profit requirements of the manufacturer.  I have a harder time seeing them beat the economics of a used sawmill, but theoretically it’s possible.

As I think through this example though, I’m reminding myself of how much difference there is between bolting together a prototype and getting it to work for long enough to shoot a video, and refining and debugging it into a reliable, day-in-day-out tool.  In one video, the founder talks about their sawmill while holding a tiny piece of slabwood that was cut on the machine.  To my mind the video would have been at least 10 times more impressive if he shot the video in front of the stack of boards that would come out of just one tree fully milled, and another 3 times more impressive if shot in front of a pile of 2000 board feet cut in the first full day of operation.

There’s a romance vs. practicality thread in this conversation, and I very much understand the romance part.  I really love working in the woods with a chainsaw and a tractor, and how much cooler would it be if you could saw out your own boards? A lot cooler, and I’m not the only one to think so – when I was a kid, in the back of the barn there was a pile of metal, some old flat belts, and a few giant sawblades; the guts of a used sawmill that witnessed to my grandfather and my dad scheming to build their own sawmill to mill timber from the surrounding 100 acres of land.  Life being what it is, the sawmill never got built, and the parts got sold to some other aspirants.  But that’s in part because there was a better solution – in Maine there are a number of guys who will roll into your yard with a portable band mill and cut your logs into lumber for something around 25 cents per board foot.    If the sustainable harvest of the land is 1/2 cord per acre per year, and 1/3 of each cord ends up as boards (which seems way too high given crooked stuff, tops, air, and slab) that’s about 25 thousand board feet of sustainable lumber per year from those 100 acres.  The productivity of those mills is probably in the neighborhood of 3000 board feet per day, so even if we were maximally harvesting the land, the saw would be in use for less than 2 weeks out of the year.  Now if we’ve got cash to burn and we’re just hobbying around, then anything goes.  But in a resource-constrained world, capital efficiency matters, and the allocation of valuable equipment, knowledge, and time is worth thinking about.  It makes sense to hire a guy who uses his mill every day to come and saw the logs; it could make sense to rent a mill and saw them yourself (if that option were available), it could make sense to buy or build a sawmill and go into business milling other peoples’ timber, but as cool as it would be to have my own sawmill, it’s hard for me to argue that it makes practical sense to own a sawmill that sits in the shed 50 weeks out of the year.

Another example: I see that OSE has a design for a PCB mill at the prototyping stage. As it happens, in college my friends Holly and JD built a circuit board milling machine for their 6.111 project.  They had a great time doing their own machining, and even cast their own aluminum pillow blocks (!)  It was pretty awesome, and they eventually got it to work (at least enough for checkoff).  But as far as I can remember, they never actually used any circuit boards they made on the thing, and pretty quickly it got broken down and scavenged.  And PCB milling is cool, but I’m not even sure that it makes sense, when superior PCBs made by conventional manufacturing can be delivered in hours, and  and there are shops that specialize in hobbyist boards at a relatively low cost and deliver in just a few days. Nor does it make sense in a mad-Max context, unless you also have the capability of producing copper foil, glass fiber cloth, and epoxy resin (or decent substitutes), not to mention the ability to make practical resistors, capacitors, and microprocessors to populate them with.  While we’re on the topic of PCBs, at one point in college my housemates and I got the crazy idea that we were going to etch our own circuit boards, and we foolishly took delivery of an old wet processing machine and a drum of nasty ferric chloride in our basement.  Needless to say, we never etched any boards, and the net effect was that we eventually had to get rid of a big nasty contraption and a barrel of toxic waste. Advanced Circuits and their kin are just too good at what they do.

I would argue that PCB mills are in a category of things that get made or bought because they are cool, but rarely get used.  And a very basic problem with a lot of this stuff is that it is way more fun to build new stuff than it is to maintain and manage it. It also seems to be way more attractive to fund the creation of something new than to keep it going.  At one point I needed to cut some fancy 3D shapes in a soft material, and I was fortunate to be able to use the CNC router at a FabLab in Boston.  The tool was in rough shape, and we spent at least half our time there cleaning and repairing it to get it to where it would run, which was fine with us – at least we had access.  But when we went to try and use it again a few months later, the availability had gone from 1 night per week to zero – because the host organization didn’t have funding to staff the space.

We did our own thought experiment in open source hardware a couple years back when we started thinking about how to build an affordable, high-quality small home wind turbine.  We did a bit of research and found that there was already a sophisticated, more-or-less open source design, promulgated by a respected designer in Scotland named Hugh Piggott (he charges a few bucks for the printed book of plans; not sure if that counts).  His turbines are solid machines, but I would argue that they are at the upper end of what highly-skilled generalist ‘maker’ types can be expected to build, and to make a small Piggott turbine of  a few hundred watts output would take a person many days. There are also some compromises in the design that are inherent in the low-res garage fabrication techniques available to most amateurs, especially relating to the efficiency of the airfoils and the alternator.   To do better than available turbines on cost and performance, we would have to take maximum advantage of all the tools of modern manufacturing, and to do that would require capital – at the very least, we would need to invest in design-specific molds and tooling capable of producing high-performance parts at a reasonable cost.  And it’s much easier to attract capital when there is  strong prospect for financial returns deriving from exclusive use of the associated intellectual property – the investing world is not yet entirely familiar with the open source hardware concept.  So while we considered an open-source project, given the practical realities of manufacturing,  we launched Pika Energy with a classical startup approach.

It’s surely not a coincidence that much of the action in open-source hardware is in electronics, rapid prototyping, and other areas (e.g OSE’s metal fab) that by their nature have low upfront/tooling costs, and I’ve previously discussed the limitations of rapid prototyping as compared to conventional manufacturing on this blog. I have a suspicion that weld-your-own tractors are going to face many of the same challenges, and I would be extremely surprised if an open source solar turbine eclipses the substantial and as-yet unsuccessful commercial efforts to date to produce motive power more cost-effectively than PV and conventional sources.  As open-source hardware develops, it will be really interesting to see if these efforts can build truly industry-leading products that address the needs of mass markets.  If they do, I predict it will involve harnessing modern manufacturing much more intensively than efforts to date. And I have a suspicion that the biggest challenges to be overcome will be organizational, rather than technical.