Making An Earthen Floor
by Joshua Alden/Fergus Reilly
NOTE From Charmaine: This article is a ten page email provided by Joshua, about the earthen floor of his strawbale home.
FROM 2002- Starting last Saturday, we installed the first two layers of our earthen floor. Here's how it went, and here's what we found out.
I prepped the floor by laying cinder blocks for a hollow-core passive solar floor. There were ventilation trenches at each end without cinder blocks in them. I roofed over these with pieces of one-by scrap pine. I wasn't meticulous about levelling the gravel under the blocks when I laid them, though I did give it a good eyeball and a couple of checks with the water level. So I knew that in one place the floor would be 1.5 inches thick, and it would thicken to four inches about twenty-five feet away from there.
I started by experimenting with different clay/sand mixes, since I'd been told by people with experience that every clay is different. The best batches were the 1:1 mix and the 2clay:1sand mix. Athena and Bill Steen recommend the highest clay content which won't crack, so I went with 1:1, but when I was mixing I erred to the side of more clay, so the end result is probably around 6clay:5sand. We added no lime, cement or straw. We mixed with a gasoline-powered mortar mixer. (It had paddles, as opposed to a cement mixer, which doesn't.)
The piled up earth shown here is from the floor building process of Peter (aka Fergus), his wife Trijntje in Australia.
We discovered quickly that how we put the materials in was important. Our clay was closer to wet than dry, so I broke it up into little chunks by scraping the big pile of clay with a hoe. I then shoveled these little chunks into the mixer. I recommend the scrape-and-shovel method highly.
I could have shoveled out direct, but it would have been harder to shovel, harder to get consistently sized chunks, and taken longer to break up in the mortar mixer. I did not lose any time by scraping; it's quick and easy. We did not try to screen the clay; there weren't many pebbles in it, and I could think of no easy way to screen them out. I knew that they would do no harm in the floor, and might marginally improve it, by being non-expansive.
We added some sand first. Our stucco guy suggested this, pointing out that the sand would scrape against the clay and help break the chunks down into slurry. He was right. So we threw in ten shovels of sand, and then some water. We learned quickly that an amount of water which looks like way too little water can easily be too much, so we were careful not to add too much. We then threw in ten shovels of clay chunks and let the mixer mix.
Then we threw in equal amounts of clay and sand until we had the amount we wanted. We decanted some of it into a wheelbarrow, which we used to get the mixture inside the house. We discovered that leaving some of the mixture in the mixer helped to start the next batch going, so we kept the batches large and decanted only part of each batch. We mixed it pretty dry. The consistency was drier than what you'd want for plastering; it was of a consistency which would be hard to move and mix by hand. It was too stiff to trowel, except for surface detail work, like floating.
A Modern Approach to an Ancient Practice
by Sukita Reay Crimmel and James Thomson
For most of human history, people have lived in durable, comfortable buildings made from natural materials such as soil, sand, rocks, and fiber. All over the globe, these ancient traditions persist; a quarter to a third of the world's population today lives in houses built partially or entirely of earth. Conventional Western building techniques using industrial materials may save time and create efficiencies, but these perceived savings come at considerable financial and environmental cost.
As well as boasting a unique and beautiful aesthetic, natural building techniques are accessible, affordable, and non-toxic. Earthen Floors: A Modern Approach to an Ancient Practice is the first comprehensive, fully illustrated manual covering the history, use, and maintenance of this attractive, practical flooring option. This detailed, fully illustrated guide explains every part of the process, including:
Sourcing and harvesting materials
Preparing the subfloor
Pouring, finishing, and sealing the floor
Living with and maintaining your earthen floor
Because information on creating quality earthen floors was not previously widely available, there have been some negative experiences. Drawing on the combined knowledge of the most qualified earthen floor practitioners, as well their own substantial experience, the authors deliver the definitive resource for this exciting technique, perfect for everyone from the novice to veteran builder. Paperback / 250 pages. New Society Publishers. 2014.
We had laid out radiant floor tubing and weighted it down with cinder blocks, so we put down long pieces of two-by pine to protect the tubing as we wheeled the mixture in. When we got the wheelbarrow to where we wanted it, we poured the mixture out and spread it around. Later we discovered that it was just as fast, and more accurate, to pick up handfuls out of the wheelbarrow and plop them down where they were needed. After plopping them into place, we used our hands to press the mixture down and spread it around. I found that a sort of quick slapping motion helped to liquefy the mixture enough to make it meld with the adjacent mixture, and self-level to a small extent.
Our objective at this point was to cover the tubing and not much more, so where the floor would be only 1.5 inches thick we just barely covered the tubing, leaving it visible in places. Where it was going to be thicker we covered the tubing completely. We had marks at the correct height all the way around the wall, but we didn't refer to them much with the first layer. Mixing and laying the floor as we went, it took two of us about nine hours to lay the first layer. The area is about 600 square feet.
As we laid the mixture, we levelled it by eye with our hands, and then ran our fingertips across the surface to make it uneven, for a better mechanical bond with the next layer. The weather was intermittently cloudy. It was not very humid, and not very dry, with the temperature usually in the sixties (F). The first layer dried quickly; within a few hours it was possible to walk on it without leaving much of an imprint, and we didn't sink in at all (I weigh about 190 lbs and have ordinary-sized feet).
I hypothesize that the cinder blocks, which were dry, sucked the moisture out of the clay mixture. The areas which were over packed gravel seemed to harden more slowly. After my helper left, I just mixed and laid the mixture aside without spreading it for a few hours. The mortar mixer was a rental, and I wanted as much use out of it as I could get. This turned out to be a good plan, because it wouldn't start on the next day, so I lost half a day of mixing. On Monday I laid about 450 square feet of floor, about two inches thick. I did it by myself, and I was feeling it in my arms and lower back when I was done.
I ran out of mixture after a couple of hours on Tuesday, so I started mixing by hand after that to get the layer done. I found that I could mix and lay a wheelbarrow-full of mixture in just over thirty minutes. I hand-mixed it in the wheelbarrow, using a leaf-shaped hoe to do the mixing. I found that it was important not to add too much water, as before, and to have more sand, then add all the clay gradually, and then dry the mixture out with the final amounts of sand. I also found that no amount of mixing would get every chunk of clay. I had to get in there with my hands, search around for the final chunks, smash them by hand, and mix them into the rest of the mixture. This also gave me the opportunity to toss the larger stones I came across.
I found that if I was willing to work hard in the mixing, I could mix it just as dry as the mortar mixer did, and only take about twice as long to do it. Working a few hours a day by myself, I laid about 150 square feet, two inches thick, in two days. From Monday onward the weather was rainy and cool. This layer was a couple of inches thick, and I was laying it on damp earthen floor, so it dried much more slowly than the first layer. Nevertheless, by Thursday I was able to walk on the material which I laid on Tuesday without sinking in. I walked all over the surface, which was smooth from my efforts to level it, and left footprints everywhere, thereby roughening the surface and preparing it for the final layer.
I noticed as I did that something about the size of a large squirrel had been walking around on it during the night, leaving neat little claw marks. We pulled some lines taut to the marks on the walls, and I discovered that I had things pretty level by eye; the remaining levelling would be easy. As the floor dried, it cracked in a few spots. I walked heavily on those spots and the cracks went away. Some of the cracks followed the edge of the pine scrap which I had laid down to cover the ventilation trenches. I'm hoping that now that the floor is thicker, we won't have so much of that cracking. So far, so good! Warm weather is forecast, so we'll see how fast it dries.
Around 05/25, I finished laying the lower two layers of my earthen floor. The total thickness at that point ranged from an inch to 2.5 inches. I let it sit and dry until 06/03, when I started laying the final layer, which ranged from half an inch to two inches, depending on what was necessary to bring everything up level. Between 05/25 and 06/03 we had warm, dry weather (for this area) and the floor was almost completely dry.
From 06/03 to 06/14 I laid the final floor layer. I was working mostly alone, with some help mixing clay and sand for the floor mixture. The work was not terribly demanding physically, though I had to be careful not strain my lower back on long days. I'm physically fit, with a strong back, and I know how to lift properly, so it's important to note that I was feeling my efforts by the end of the day. I laid about twelve or thirteen square feet per hour. Unlike the previous layers, which were much faster, I had to be finicky about getting things as level and smooth as possible.
I didn't keep close count, but I'd say that the final layer took about thirty wheelbarrows-full of mixture. Each wheelbarrow mixture has, at a rough guess, twenty or twenty-two shovels-full of clay, and an equal amount of sand.
This was my method:
I tied a length of baling twine to a cinder block and gave the block a few turns to wrap the twine around a few corners. Then I ran the twine to another block and did the same thing. By turning the blocks I was able to take up slack or let it out for length. I set each block close to opposite walls, so that the string was taut, and sighted along the string to put the plane of the string just above the marks on the walls. I moistened the previous layer by splashing some water down and slopping it around.
Then I put the next layer down on the first one and carefully flattened it so that it was just under the level of the twine (if I brought it right up to the twine, it pushed against the string and altered its position). As I laid the mixture, I came across occasional small stones in the mix (it was too moist to screen when we were mixing). When I did, I pushed them down against the previous layer, or if I was in a thin area, tossed them out the nearest opening in the wall.
As I laid each strip of mixture, I scalloped the edges with my fingers. That helped the strip bond with the next strip, especially when the next strip happened the next day. When I had the first section laid, all against the west wall, I moved both blocks a couple of feet further east, checked the position of the twine, and laid more mixture in the same fashion. Then I used my aluminum level (see below) and a plaster trowel to work the mixture flat and smooth. In this way, I laid all the way across the house, which is about 600 square feet.
On 06/15 and 06/16, I did the final levelling and surface finishing. I used a four-foot aluminum I-bar level to screed the moist surface, which trimmed the heights and pointed out the lows. This screeding produced a fine, grainy soil mixture, which I used to fill in the low points. We also brushed it into the cracks which had formed, and that helped fill those up when I poured water over them and swished it around. The screeding also picked small stones out of the clay. I threw these stones out through the nearest openings and filled the little divots where they had been with the same grainy mixture. I used a small amount of water and the trowel to smooth the surface. I worked across the entire floor surface that way in about fourteen hours. That also included building a small ramp out of the same mixture, at the side door.
Up until I built the ramp, I was just trying to get the mixture as flat and level as possible, so the versatility of the material did not really show. But I started to realize as I built that ramp how much fun it must be to make a cob house, even if it is labor intensive. It's incredibly flexible; you can do just about whatever you want, and unlike lime and cement, you can shape it with your hands, for as long as you like. I managed to get things pretty level. It's not as smooth and planar as self-levelling concrete, or a newly-installed professional wood floor, but it's pretty good. If I cared enough to delay the building process, I could go over it again and get it even closer, but I'm looking at diminishing returns and at this point I'm happy with what I've got.
The process of laying the second layer and then troweling it smooth introduce quite a bit of moisture to the floor. I'm now waiting for it to dry. If I'm careful now, I can walk on it without leaving any impressions; it's still cool to the touch, and definitely not dry. It's been five days since I finished, and it's drying much more slowly than it did last time. I have a fan playing across it all the time, and I'm using a smaller fan to drive warm air through the ventilation trenches, which has to be helping dry the floor from underneath.
The floor around the pipe which feeds the hot air into the ventilation trenches is now completely dry, but only for a few inches from the pipe. I'm starting to think that it will take too long this way for our current plans, so I'm hoping to fire up the heater in a few days and start putting hot water through the radiant tubing in the slab, to help things dry out. I'll post more as it happens.
My last account went out on 06/22, when the final layer had been down for a few days, and I was waiting for the floor to dry. We had a long stretch of very humid weather, so the floor dried painfully slowly. I was not able to get the radiant heat hooked up until after it looked like it would be useful, so I did not end of driving heat through the floor. However, I could not wait for the floor to dry; I needed to keep building. So, I walked barefoot, used a board under the feet of the ladder, and tried not to drop tools as I wired the house and put down wooden flooring in the loft. In a perfect world, I would not have had to do that; I did drop some tools, with resulting small divots in the floor, and the surface did roughen up somewhat after my lovely smooth trowelling job. But financial pressures did not permit a delay.
It turned out to be very easy to patch the little divots, and the patches dried very quickly indeed, compared to the floor as a whole. I did encourage drying by driving air through the hollow-core blocks. I had a small 1500-watt electrical heater with a fan, and I stuck it up against the 4-inch tube which feeds the air channels. When I was working there, I turned the heat on, and overnight I just left the fan running. Judging by the way the floor near it dried, it greatly accelerated the drying. Down in the southwest, these floors probably dry in no time.
There was one snag which worried me. I had a layer of cinder blocks for a hollow-core passive solar floor, and in order to distribute the air from the source point to the blocks, there were three trenches which I roofed over with scrap pine. That was all right in most areas, but I had to go pretty thin with the earthen floor in the kitchen (1.5 inches). The cracking during drying was minimal, but as I walked around on the floor, some of those boards flexed slightly, with resultant cracks, some of them quite large with visible disintegration of the floor material into little crumbly bits. I took up the worst sections and re-fitted the boards so that they did not move as much. The new material dried quickly because it was thin. However, it cracked again as it dried. I considered various permanent remedies, but they all added water to the floor, and I wanted it to dry. So I just crossed my fingers and kept working.
The floor was finally dry enough to put on the first coat of penetrating oil on 07/20. I used Natural Choice Penetrating Oil Sealer #5. What a difference! It was immediately much tougher, seemingly by orders of magnitude. The floor soaked up that oil almost like a sponge takes water; it takes a lot of oil for the first coat, but the subsequent coats took much less. Around 07/28 it was dry enough to put down a second coat, and I did. Around then I moved the pressure tank outside and installed about twenty square feet in that area. The mix was just as wet as the rest of the floor was when I put it down, but it dried quickly (a week or so) compared to the floor as a whole, despite being two inches thick, and over gravel rather than over cinder blocks.
While it was drying, I was working in the loft. I accidentally dropped a piece of pine, 2x16x16, roughly. It landed on its corner (of course) about nine feet down, on the floor. I went down, fearing the worst, and found that it was difficult to find the dimple where it hit. It's there, but you'd never notice it. I may try to patch it and I may not, but it's wonderful to see that it's tough even with one coat and a wet second coat. Imagine what it will be like when all three coats are on.
On 08/10, the floor had been ready for the third coat for about two days, but the oil didn't arrive until then. I oiled most of the floor, and it took a very long time to dry. The third coat might even be superfluous, and ten liters was way more than enough to coat 600 square feet on the third coat. Fortunately, I still had that section of the bathroom which needed first and second coats, so I didn't end up with lots of unused oil.
The total floor area is just under 600 square feet. It took 42.5 liters of oil to do the floor. At $85/10 liters, that's about $360, which is steeper than I had hoped, but I think it is going to be a very strong and beautiful floor. Now for the final note. Remember those bad cracks over the scrap pine? After the second coat of oil had mostly dried, I realized that the separate sections were not moving independently at the cracks anymore. The cracks had filled in almost completely during the process of swabbing the oil around, and seemed to have almost welded together. I won't be jumping up and down on it until I have the third coat in and dry, but putting my 185 lbs on it doesn't cause any movement at all. Very impressive!
It's now the end of 2002. I should note that after the oils had dried completely, we waxed the floor to give it a bit of protection. The floor has held up pretty well under constant use. There are divots in a few places. The bathroom did the best, because there is no heavy furniture in there, and we are almost always barefoot. There's essentially no wear. In the main area, the floor needs to be protected from point loads. We put something under each leg of the chairs to distribute the weight over a greater surface area. For our rocking chair and any other chair, a braided rug was sufficient to protect the floor.
The main problem is one which I didn't anticipate: any wood which sits on the floor for a long period (several weeks) without moving at all tends to fuse to the floor. When you move the wood, it brings up about an eighth of an inch of floor, which leaves a rough patch. We've defeated this problem by putting scraps of carpet, pile side down, under the bits of wood where they hit the floor. If they're cut to size then you really don't notice them. This adaptation works for us because we don't move our seating much; it's a small house and seating is pretty well fixed. But it would not work for many people. I doubt that this problem happens with all earthen floors. Perhaps Bill Steen could address this.
The end result for us is that we intend to lay some tile, starting in the kitchen. We may do no more that the kitchen and the laundry areas. Frankly, the areas in front of the two doors are doing pretty well, because the issue isn't foot wear, but rather heavy point loads. And enough braided rug would take care of the problem, but won't work in the kitchen. -Speireag. ... SOMEWHERE IN THE WILDS OF NEW HAMPSHIRE.
Speireag (Joshua Alden) Lives in New Hampshire, is active on the straw bale elists, and has built a home using natural materials: bales, cob, clay, lime plasters, and more. Reproduced with his permission.
Hardening Earth Floors & Finishing with Linseed Oil
Adobe Floor Basics: How to Build a Dirt Cheap Floor
Earthen Floor at the Middle Path Health Awareness Retreat