Building a Cob House That Won't Burn Down
Cob is basically dirt and straw. It doesn't burn, which sounds great until you're sitting on a site at 2 AM watching a 3,000-degree wildfire push against your walls and realizing that while the cob itself won't ignite, the roof above it absolutely will if you used the wrong material. I've seen three cob homes lose their roofs to nearby brush fires in the last decade. The walls were fine. The inhabitants were fine. Every single time it was the roof that failed first. So here's how you actually do a Cob House With Fireproof Roof without ending up with a soggy pile of mud after the first heavy rain because you didn't account for thermal movement in the flashing.
Cob House With Fireproof Roof: The Actual Build
Start with the cob walls. Standard mix is roughly 40% clay soil, 40% sand, and 20% straw by volume. You want a subsoil that tests between 15 and 30 percent clay. Run the jar test if you're unsure. Pour water into a mason jar with your soil, shake it for five minutes, let it settle, and measure the layers. Anything below 10 percent clay and you're building with sand. Anything above 35 percent and you're going to get catastrophic cracking. There's no way around it. The cob walls should be finished to within six inches of the top before you even think about the roof. Leave a flat, level cap. Trowel it smooth and let it cure for at least two weeks. You can plaster the walls any time after that, but don't plaster the very top course until the roof is on and weathertight. Wet cob underneath plaster is a delamination problem waiting to happen. Now the roof. Your options are standing seam metal, clay tile, or slate. Metal is by far the most common in the cob world because it's light, it goes on fast, and it lasts forever. The issue nobody warns you about is thermal expansion. Metal roofs move. A 30-foot span of corrugated steel will expand and contract enough to work itself loose from standard nails in a single heating season. Use concealed clip systems, not exposed-fastener panels. The difference in long-term reliability is massive. I learned that the hard way on my second cob build where the exposed-fastener roof started weeping after year two because the gaskets had compressed past their recovery point.
The Flashing Problem Nobody Talks About
The detail where the roof meets the cob wall is where most cob roof failures happen. You need a drip edge that extends well past the cob face — at least three inches of overhang minimum, preferably four. Cob wicks moisture laterally. If water touches the top of the wall and isn't kicked away from the surface, you're watching the top few feet of your wall erode over three to five years. I measured it on a project in Sonoma. The client thought the dampness at the base of the north wall was a drainage issue. It was the roof overhang being one inch too short combined with a missing splash. Six months and a-calculated overhang fixed it completely. Use a rigid insulation board — XPS is fine, polyiso if you're in a cold climate — as a separation layer between the roof structure and the cob. This does two things. It breaks thermal bridging, which matters more than you'd think with thick thermal mass walls, and it gives you a clean, dry surface to flash against. Staple a butyl tape sealant strip to the bottom edge of the sheathing, then bend your flashing up behind it and down over the drip edge. Seal the vertical leg to the insulation board, not to the cob. Cob is too irregular to get a reliable seal on directly.
Get the Full Details
Counter-Intuitive Things About This Setup
People assume that because cob is thermal mass, you want maximum heat transfer from the roof into the walls. That's backwards for a fireproof roof scenario. A metal roof reflects most solar radiation. If you pair that with a cob wall that's already insulated from the structure above, you're creating a system where the thermal mass stays relatively stable and the roof sheds heat rather than conducting it. In hot climates this is the difference between a comfortable interior and an oven. In fire-prone areas it's the difference between your roof assembly staying below ignition temperature and your ceiling catching fire from radiant heat exposure. Another thing: the heavier the roof material, the more you need to think about your wall thickness. A clay tile roof on a 24-inch cob wall is totally fine. The same tile on a 16-inch wall in a high seismic zone is asking for trouble. Cob has good compressive strength but near-zero tensile strength. I've seen 16-inch walls crack along the base because the roof dead load was pushing the wythe inward during a minor earthquake. If you're using heavy roofing, go 24 inches minimum or add a structural frame inside the cob skin.
When This Approach Fails Completely
A metal roof on cob is not appropriate if you're in a area with frequent hail. I'm not saying it'll destroy the house, but replacement panels after a big hail event run about eight dollars per square foot installed, and sourcing matching metal for a repair on an odd color or profile can take months. In those climates, clay tile or a Class A rated composite that's been third-party tested for impact resistance is worth the extra labor cost. Also, if your cob walls are going to be exposed on all four sides without any eaves overhang — say you're building a cabin style with a flat or low-slope roof — the water management problem compounds. Low slope means slower water runoff, which means more opportunity for wind-driven rain to get under the metal seams. I wouldn't recommend a cob house with anything less than a 4/12 pitch unless you're willing to seal every single panel seam with lap sealant and inspect it annually. The only time I'd walk away from this approach entirely is if the site is in a very high wildfire zone and you're looking at ember exposure as the primary threat. In that case, a cob wall with a metal roof is good but not optimal. A double-wythe cob wall with an insulating airspace, a metal roof with sealed soffit vents, and non-combustible window shutters will outperform a single-wythe cob with a standard metal roof by a wide margin. The problem is cost and labor. It's significantly more expensive and takes roughly twice the wall-building time. If budget is tight and fire risk is the main concern, I'd prioritize the roof and vent sealing over doubling the wall thickness.
