This is a great discussion to have! Thanks for getting it started Jason! We could start with listing some of the main types of existing technology, calculate the costs of each off the shelf, calculate the cost of the means of producing them, and rate them for how they measure up to the OSE specifications. Then we can see which would be good to use off the shelf for now, which would be effective with open sourcing the means of producing the materials, and which would be best as a long term goal to move toward.

It looks like the sawmill was supposed to account for roofing, but it looks like a large project that wasn't finished and hasn't had a lot of work on it for a while. I wonder if a band saw mill with an open sourced automatic sharpener would be as effective/low cost as the proposed dimensional sawmill? We would have to look at the work flows and the associated cost of components for each work flow then compare them to what the cost/time would be of open sourcing the most expensive components.

There's also sheet metal over other support materials (wood trusses) for roofing, which
the GVCS set would seem to account for with the metal rolling and sawmill.
http://openfarmtech.org/wiki/Hot_Rolling Not much on it yet though, it
should probably be considered for a later stage since sheet metal is
currently kind of cheap compared to the infrastructure required to
produce it.

So it looks like the GVCS has some roof basics, but are they the best method compared to the OSE specs? The current materials seem to be low enough cost to buy off the shelf for now, as the workshop project seems to demonstrate, so the machines to produce the materials could be more of a later stage priority.

Ferrocement looks attractive from a construction and durability standpoint. The only problem I've heard of is that it takes a lot of energy to make cement, and it's only as cheap as it is because of low cost oil/NG/etc. Metal is a somewhat similar situation though. Still if it's the cheapest thing off the shelf then it will work for initial structures until OSE spec technology can be made. Can it be made with lower cost/non-metal frameworks or is the metal strength essential? What would the off the shelf costs be compared to the off the shelf metal+wood costs in the workshop example? Are those bulk buys possible and how much area would need to be covered for it to be economical? If the off the shelf cost is lower then all OSE buildings should be using it until lower cost technology/designs are made.

There are also the CEB vault roofs. It looks like more research and prototyping still needs to be done for them though to see if they can work with the Liberator's bricks or some other way of producing them.
http://openfarmtech.org/wiki/CEB_Vaults
http://openfarmtech.org/wiki/Arches,_Vaults,_and_Domes

We should also consider how insulation works for each roof type, and how much labor is involved with each type.

ferrocement.  would be my first choice.   A simple vault built up by hand by slapping concrete over plastic supported by chicken wire.

Okay, this is what its basically based on: HERE

Dome's are the way to go..here is a technique that I learned about years ago...that just seemed like a easy but solid idea
http://www.domtec.com/process.html

I just saw this house recently...that I thought was beautiful, functional, creative, easy to maintain...check it out...let me know what you guys think ?

http://twistedsifter.com/2010/02/house-inside-a-hill/

I'm not a architect by any means...but the two ideas are compelling.

@ Jason,  consider the "Cascaje" technique for your roofing: It is like ferrocement, but with less cement....

http://tibazine.wordpress.com/arquitetura-de-terra/casa-de-mirante/monthly-progress/may-2011/

@Jason well, I have to vote for @ARGHaynes idea - support some dirt and plant a garden.  The pond liner will do an awesome job of keeping the water out and the soil will keep the heat out.  Comfort is a good thing.

here is an idea for a lightweight living/green roof:
http://kevinsonger.blogspot.com/search/label/lightweight green roofs

Here's a super cool video of the stuff being used to make cover a flat roof (epdm = pond liner, same stuff):


It seems like the 45 mil epdm pond liner available online for between $0.55 to $0.65 per sq ft, shipped. For example, here or here. You can also get stuff  specifically for roofing, which I've seen as 60 mils for $1.05/sq ft + shipping.

Of course EPDM probably isn't something we could make ourselves (or is it?), but I'm all down for pragmatic solutions. Anything to get the job done. We could have some EPDM roofing designs in our repertoire of roofing solutions.

Part of localization means using what is local, epdm is local to where I live.  The question at present is not about every possible roof solution, but about Jason's roof solution.  I covered my chicken coop with a metal roof and remember it being more expensive than I anticipated, plus there is no insulation properties with plain tin.

Making the cement would also be a challenge, for that matter. Although it is much more readily available.

Here's something that could be done at FeF: ferrocement vaulted channels.  It is a very elegant roofing solution for modular housing.  See the attached for more info.

they are arcs already, and they can sit right on the walls.  The support beams go on the ends, perpendicular to the channels, like the top of a wall.

But, you could easily modify these shapes to fit anything, just adjust your form.  I think it would match up very nice with CEB modular housing.

Now, combine this concept with a laminated ferro cement technique, and you could have roof channels that are 1" or less thick, lightweight, easily produced, and extremely inexpensive.

The tape sounds tolerable, I do suggest a raised seam - fasten the bottom edge of one piece to a handspan from the edge of the bottom side of the other, then fold the extra length over and fasten it to the back of the shorter.  This can be supported by some boards or bricks or something to keep it upright, and the pattern should make it even less likely to have a leak.

@Jason
On the "$50 underground home" videos the guy says that flat roofs or roofs with flat parts where water can sit, will always end up leaking eventually. Have you planned to make sure that your roof is going to shed the water effectively from all areas? This principle should also probably be researched and documented on the wiki page for roofs if it is a valid concern.

Let me give everyone an update since I last posted. I've got the first section of 2x6s up. I went with 2x6s at 16 inch centers. The area being spanned by them is 83". I also changed the way I'm doing the joints, etc. Let me give a few illustrations.





I think the idea of putting some dirt on top and planting stuff is a pretty good one. We'll see if I do it or not. I think it should be strong enough though.

As far as making it so water doesn't pool...that is a good point John. I appreciate your feedback. I think I'm going to put drywall shims under the center of the span to give the whole thing a slight convex shape. That should eliminate pooling (though I am using pond liner...if anything could handle pooling I think this material would :P).

Also, I'm going with the Type I joint above, not the Type 2. The 6" wide tape is much more expensive than the 3" wide tape.



For the 2x6 structure, aside from being stronger and just better in a
number of ways than the previous way I was doing it, I also don't have
to buy the nuts and bolts ($40 in hardware) or the butyl caulking
sealant ($20).

Calculations for 6" of water over the roof (dimensions 39' x 7') shown here yield 4 tons.
The same calculations for soil at 100 lb/cubic foot here yield 6.4 tons.

Yes, I do think it could hold that much. Assuming 10 tons, that would be about 666 lbs load on each spanning 2x6 (assume 700 lbs worse case). This is absolute worse case, and doesn't count the double think 2x6s every 8', and it also doesn't count the reinforcement by plywood on top.

Also check out this calculator:


Anyway, I've already changed my plan a bit. It won't be just an even 6" of soil. More details to follow.

Well if there's going to be an open source robotic arm, and if the Liberator can be modified to press roof tiles, then the arm can be programmed to lay the tiles.

Here's an example

Making comments to the entire thread as I read through it...

I agree, a roof is perhaps the most critical problem. You can make strawbale walls, earthblock walls, you can get by with a dirt floor, but without a roof things aren't much use. From some of my own readings on exactly this problem and trying to come up with my own solutions:
- Light steel framing is strong, lighter than wood for the strength, and apparently cheaper. One downside is thermal conductivity in colder climates, another may be fire risks... wood survives better than thin steel at a certain heat level, the outside charrs. This isn't a veto trait, just something to consider. Superinsulation strategies (ie paying particular attention to the connection points, since heat gain/loss flows in 3 dimensions, and in such a roof would be flowing through the steel roof members... so you insulate the roof members at every attachment point to minimize this) may help mitigate this.
- Ferrocement NOT necessarily a great idea if you ever have any risk of earthquake. (and even parts of the country that don't have earthquakes are starting to have them due to natural gas/oil extraction fracking going on, so this is of concern in places where even it hasn't been for centuries) Big heavy roofs falling on you tend to = death in every third world country, especially if the engineering sophstication may be below state of the art. You need very careful engineering, a good reason to have a heavy roof (protection or underground housing may be good reasons however) or maybe the use of ferrocement to be more like a sealant, mortar, or way to tie together individual blocks so individual blocks cant fall out or be knocked out of an arch/dome making it collapse being over lighter weight materials. It's also damn heavy so youre talking a nontrivial way to raise the roof unless it's built in place somehow.
- Corrugated steel sheets is used for roofs all over the third world due to cheapness, durability, and such, the bigger question is how to support to and how to highly insulate it. It's ugly but one of my personal favorites for cheapness, and can cover over things like strawbales.
- Wood OR homemade light steel trusses are doable, but design software would be needed. I'd like to saw my own dimensional lumber, and make my own roof truss... but it's not very efficient without design software, either overbuilt, weak in general, or not strong in all directions it needs to be for instance.
- Domes and arches are my personal favorite. The use of slightly arched earthblocks could be one way to do it, you could make arches and domes out of strawbales too - some others have tried, with varying results. It's an idea I personally like.
- Remember it's not just cheap materials, you also have to build it possibly by yourself. Already stated but one thing favoring any lighter roof over a heavy one. Heavy ones require tools that can be pricey as well killing the money saved.
- Dont forget the insulation. Roof insulation is among the most important in the whole house since heat is lost that way in the winter. Strawbales is one way to do it if you can support their weight well, another idea would be ricehulls which can freefill available space to any thickness desired and are considered a waste product/free or very cheap.
- Dont forget thermal gain in the summer. Heat gained through the roof adds tremendously to either AC load or simple uncomfortableness if it's an uncooled or just vented house. Reflective or white roofs are better, anything dark like conventional shingles might hit 180 degrees in summer from solar gain and transfer much of that into the house. In some environments with a wide delta like the desert (where during day it's too hot for comfort but at night it's too cold) this can be used to advantage if you have a thickness of thermal mass so that the sun beating on the roof all day takes about 12 hours to "go through" and then radiates down during nighttime, this also stabilizes temperature in general. This can be one reason to justify a heavy roof. Earthblocks provide this 'time lag' thermal transfer just fine btw.
- The amount of work it takes DOES matter. Free that takes hundreds of hours vs very inexpensive but automated and taking almost no time... it's been said that earth friendly housing is often a project for either the idle rich or the idle poor. The working poor don't have enough hours in the day to say make shingles one at a time by crushing cans and such, although i'll admit such shingles are clever.

So I guess we have four main areas to consider, just to restate:
- Framing. Whether wood, metal, or somehow not needing a framing (monolith concrete slabs perhaps/thick enough or of a design that is self supporting) you still need design software to make the trusses efficiently. This seems to be a critical problem to me if you don't want to buy a commercial truss or roofing system. Eventually some type of bioplastic may be usable if it could be made fire resistant enough.
- Insulation. Fiberglass is toxic like asbestos to the lungs/having fibers the same size. It is lightweight but not recommended for that matter. Reclaimed styrofoam IS often available secondhand for very affordable prices if you know where to look, it's also more efficient per inch, it's what I will probably use if I can get it. Strawbales are a bit heavy but cheap for the insulation value, they just have to be covered with something and supported with strong framing of some sort. Loose straw with borax to give it fire resistance is an option for less or thicker than strawbales are. A Truss CAD is still needed or premade scalable designs in several sizes. Natural fibers are an option - shredded denim for instance, though I don't know of a way to buy that bulk and make your own natural insulations that way/not sure where one gets such a thing. Ricehulls. Use of unsafe improvised insulations like say styrofoam packing peanuts is probably NOT recommended due to fire hazards. (could borax fix this like it does with straw?) Thermal mass in the roof (primarily earth, possibly cement, rock, or deliberate water tanks or something) can have benefits but by definition adds weight.
- Covering. Whether shingles, rolled steel, or a waterproof plaster, it should also resist things like hail. Someone mentioned cement shingles, i'd fear them cracking on hail if they are thin enough to be light and cheap.
- The Building Process. What tools are required to put it up, how much work, etc.

Now commenting on others posts:

I'm not sure why wood/metal roofs are "not long term solutions", metal maybe lasts 25-30 years before rusting bad which is probably long enough for most people. There are wood houses older than the united states. It's possible that certain treatments, polymer layers added inexpensively, or reprocessing of the metal when it gets really rusty at some point could make it usable much longer term.

The $1/sq ft cement roof is interesting and seems light for concrete but but it's uninsulated, which isn't as much use to anyone outside seasonably warm climates like on their house or something... the entire cost of the entire roofing system, all inclusive, including modifications to the house under it and the tools to put it up needs to be taken into account.

I'm not anti-cement just pointing out some problems... For "cement" style roofs consider light weight aggregates like pumicecrete is an option to have it insulated and everything, then it has to be thicker to have enough insulation thickness however. It will still weigh many tons. I still would be nervous in an earthquake unless it is very very solidly built, and even then. I remember reading a site talking of low cost nontoxic housing and being depressed when I found out that even low cost pumicecrete would cost tens of thousands of dollars. Stabilized earth is what i'd probably go with if I had to and i'd think beforehand what insulation I wanted to use and how to support the weight.

Flat roofs arent good but a 'shed' roof (like a flat roof just at an angle) is probably the cheapest and simplest roof style if you dont mind one side being higher than the other. It's simpler than anything with a peak. If not an arch or a dome i'd want to use a shed roof for cost reasons.

Green roofs + + + especially on any limited landspace, it makes sense to use all the land you can for something useful. Houses take up alot of space on a smaller lot or subacreage location. My own preference is for an underground house basically under a greenhouse personally for best use of space.

Poured earth (especially if stabilized, i'd think you can do that?) might be an interesting way to make organic forms and nonangular shapes which might open up new possibilities at a low cost... I remember seeing some neat nonangular designs intended for cement, that I thought would be neat if you could do them lower cost with pumicecrete or lowest cost with something earth based. If you could build it out of bricks or at least sections somehow that could ease the building process as well.

When thinking of wood, also include bamboo as an option...


The "CO2 output of cement" construction should be irrelevant and i'm surprised it's even the discussion, carbon dioxide is not deadly toxic waste, demands to reduce it is a globalist scam to tax all human activity while ignoring real problems. Pollution potential of methods is a concern, carbon dioxide is not a pollutant, it's what plants breathe and they pump it into greenhouses to make plants healthier.

Energy input IS a concern, concrete is very energy intensive. It might be okay when "subsidized" with cheap oil energy inputs for low cost housing, but in terms of long term sustainability without cheap energy it is not sustainable.

Solutions that work now (ie what is more or less subsidized energy used to make cement, I mean you cant make it yourself for cheap) are not the same as options which are best "ultimately", and the solutions in different environments will be different. Therefore multiple solutions are going to be expected and needed, the solution for the desert is different than one in northern canada. But lets just figure out solutions that work now and worry about ultimate long term perfect sustainability later...

Did the "ideal $1 a square foot" roof ever get invented and used on here? Just curious, because it's been a few years, and roofs haven't gotten cheaper.