Building the Cob and Tadelakt Outdoor Fireplace – Homemade Tadelakt!

FP 00_starwatermark

It has been a dream of ours to have an outdoor fireplace for some time. We toyed with the idea of building one using an old propane tank or of refurbishing an old cast iron stove. In the end however, we decided to stick with what we know and build one from cob. We had a lot of left over clay soil (Burkeville Ready Mix) from building our pizza oven and tandoor so that is what we used to build the fireplace.

Hopefully some of you are inspired to build one of your own. Our family, friends and pets love sitting in front of the open fire and we regularly gather there to drink hot cocoa or a beer and to roast chestnuts and marshmallows. Below are the step by step instructions of how we went about building our fireplace.


FP 000_starwatermark
The tadelakt catching the moonlight.



The measurements of our fireplace were based on the detailed explanations that can be found in Adam Weismann and Katy Bryce’s book, Building with Cob. Ours is a Rumford style fireplace, characterized by a forward leaning back wall and angled sidewalls to direct heat forward rather than up through the chimney. This blog gives the dimensions that we used but to respect copyright laws we refer you to Weismann’s book (referenced at the end) to learn about the ratios and rationale on which our fireplace is based, as well as cob construction in general.

When we started we really didn’t have a solid vision of how our fireplace would turn out. We started by working out the inner structure and the outer form just naturally took shape around it over time. At many points during the construction we had significant doubts that we would end up with a working and pleasing fireplace! Having built this one, we now see how much room for creativity and art there is in a cob fireplace’s construction. With this one being our first attempt however, we kept the design pretty simple. We are already brainstorming ideas for our next fireplace project!

We first needed a site. We chose a previously unused piece of our land, well sheltered from the East and North by trees and open to the West and South for sun. It is close enough to the house for easy trips to the loo but far enough away that it feels very private and outdoorsy.

The inner foot print of our fireplace’s firebox is 3’ wide in front, 18” wide in the back, and 12” deep forming the shape of an isosceles trapezoid. With this basic shape in mind we dug the foundation giving ourselves roughly 8” extra girth on the back and two sides for the walls. Our foundation also extended out to the sides because we planned to flank our fireplace in cob benches. We filled the foundation hole with sharp gravel. You will need your foundation to extend below the frost line for your climate. The foundation provides drainage beneath the fireplace and, we hope, will also prevent it from sinking too quickly or unevenly.


FP 0_starwatermark
Schematic of foundation with overlayed firebox outline.


FP 1_starwatermark
Filling the foundation with gravel.

We arranged rocks on top of the foundation, using boulders to support the benches, concrete cinder blocks for the base of the back wall, and small rocks to line the front. The holes of the cinder blocks were filled with gravel. When using concrete as a cob foundation, be aware that it exerts a strong suction force and can wick water from the ground up in to the cob wall. To avoid cob failure, ensure that your concrete is either non-continuous (ie composed of clustered pieces rather than a solid poor) or that you incorporate a water break (such as a stone layer) between the concrete and the cob to prevent wicking.


FP 2_starwatermark


We built a small cob wall over the front stones. The purpose of the cob is to raise up the floor of the firebox so as to house sufficient insulation beneath the bricks. It also provides a surface on which to rest the front-most floor bricks of the firebox. This little cob wall should be as level as possible and is made using regular cob (ours was 1 part Burkeville Readi-Mix, 1 part sand, with lots of straw). Our cob mini wall was tall enough to conceal the insulation as outlined below. We wanted our firebox close to the ground to keep our feet warm however you may choose to raise your firebox up and build a cob or stone hearth for sitting on. The hearth could also be sculpted to create a storage and drying bay for wood. So many possibilities!


FP 3_starwatermark


The area enclosed by the cob mini-wall and the back wall was filled first with a layer of dry sand (roughly 1” thick), and then a thick layer of wood ash. We put down sand first to reduce the amount of wood ash that would disappear down between the gravel. We only insulated the area that would be directly beneath the firebox foot print, leaving the areas of the foundation that would support the side walls empty. Wood ash is a terrific insulator. Sand and gravel absorb and store heat and so are by contrast terrible insulators.


FP 4_starwatermark
Jen layers wood ash over dry sand.


We nested wine bottles in to the wood ash and filled the spaces in between with wood ash. The wine bottles trap air spaces, also a terrific insulator.


FP 5_starwatermark
These empty wine bottles were covered in more wood ash prior to the final layer of dry sand.


The bottle/woodash foundation was then covered in a layer of sand leveled in all planes to the height of the cob mini wall. It is important that this part, along with the cob mini wall, be level otherwise the entire fireplace will lean one way or another. The empty areas on either side of the fireplace footprint were filled with gravel. Weismann’s book describes an alternate way to build the foundation.


FP 6_starwatermark


For this project we used cheap, low density firebricks since we couldn’t think of a good reason to have the bricks retain heat as with the cooking ovens. We arranged bricks to form the base of the firebox, bringing them right up to the outer edge of the cob mini wall.


FP 7_starwatermark


And then we built up the brick walls. Our walls are three bricks high, dry stacked. LEARNING POINT! When fires burn in these fireplaces, wood that leans taller than the brick back wall erodes up the firecob back wall causing quicker failure. Build your bricks tall enough that your wood will rarely, if ever, be so tall as to lean against the cob. We cut the bricks with a stone cutting wheel mounted on an angle grinder to get the angles that we needed. Google Sketch-Up worked great for sorting out the angles precisely!


FP 8_starwatermark


FP 9_starwatermark


Then the cobbing began. We used regular cob: one part clay rich soil (yes, still Burkeville Readi-mix!) and one part coarse sand/small pebbles with oodles of straw. We packed in behind the bricks, working up and over the top to lock them in to the place. We left about ½” of the top of the bricks, towards the inside of the fireplace, uncobbed, creating a little brick ledge on which the fire cob would later be built. We had problems with the brick wall toppling so we worked quickly to build a sloppy wall to lock the bricks in to the place. We then went back to tidy it up once it was stable.

LEARNING POINT! We found that leaving a thicker lower ledge (slightly deeper than the thickness of the bricks) allows for a thicker firecob back wall which slows the rate at which the back wall cracks and fails. With this fireplace, we had to repair the back wall after 1 year of use. The new thicker firecob wall is holding up much better so far.




Sloppy quick wall.


FP 10a_starwatermark
On second pass we tidied it up.


We decided that we wanted the front opening of our fireplace to be arched. It is also possible to build a square opening using a lintel as outlined in Weismann’s book. We initially planned to make a round arch but after drawing out a few different designs we liked a slightly pointed arch the best. Our front opening is 3’ high at the peak of the arch. The opening’s sides run vertically for about a foot before curving inward to form the arch. There are countless ways this part can be designed. We first sketched out the shape we wanted our arch to be in true size on cardboard. Using this sketch as a guide, Ryan then built an arched form out of scraps of wood. Our form is shown below. It worked great.


Testing the arch fit.


The fireplace walls continued to grow up to a height of 4’ on all three sides prior to putting the arch in place.




Once the walls were at 4’, we used fire cob to build a sloped back wall that ended in a smoke shelf, 4” wide, 4’ above the firebox floor. We did this by laying a straight edge across the fireplace, resting it on the top of the side walls 4” perpendicularly out from the back wall. We then scratched a straight line in the cob of the side walls so as to connect the point where the straight edge rested at the top to the corner angle of the brick back wall at the bottom.

Then, using the little ½” brick ledge left uncobbed on the bricks, we began building up the back wall with fire cob (1 part Burkeville Readi-Mix to 4 parts coarse sand, no straw). The fire cob back wall began as ½” thick at the base, and as it grew in height so too did it grow in width and depth as it was continually built out to follow the lines scratched in the angled sidewalls. Once the fire cob back wall got to the height of the existing cob walls (4’ from the firebox floor) it was 4” proud of the bricks. Weismann’s book provides a good explanation of the purpose of the smoke shelf. It is a central element to the function of this style of fireplace. We also plastered the rear parts of the side walls in fire cob. Ultimately the entire inner surface of the fireplace will be plastered in fire cob.


FP 12_starwatermark
Forward sloping fire cob back wall.


The fire cob was quite rough and uneven, even after working it with a knife and a sponge. To create a more polished interior we decided to plaster the fire cob in a finer, more artistic fire cob plaster. We made this using spent clay that we got for free from a local pottery studio (a great place for endless supplies of pure clay in all colours imaginable!). We combined this in a wheelbarrow with roughly 7 parts sand sifted through window screen to create a gritty plaster the consistency of whipped icing. It was heavenly to work with. This was then plastered over the fire cob back wall (as well as over every surface on the interior of the fireplace up to and including the chimney).


FP 13_starwatermark
Ryan mixing the fire cob plaster. Buttery!


FP 14_starwatermark


Once the back wall was completed it was time to install the arch support form and begin building the front wall of the fireplace. The form was covered in several layers of newspaper to facilitate its removal when the time came. The thickness of the front wall was such that once it reached the level of the smoke shelf, there was a 4” gap between the smoke shelf and the interior of the front wall. Apparently the dimension of this space is important to the function of the completed fireplace. We took great care to ensure that inner walls of the fireplace were smooth and even. We incorporated a smooth rock that we uncovered when digging the fireplace foundation in to the front wall for decoration.




The front wall was brought up to level (4’) with the rest of the walls.






We continued to build the fireplace up in height, building the back wall up so as to leave the 4” smoke shelf free. As we built up, we would stop every foot or so to plaster the interior with fire cob plaster. If you build up too high it will be difficult to reach down in to the fireplace to apply the plaster. To hasten the rate at which we were able to cob, we built a little wooden form that we used to create uniform cob bricks that were the thickness of the wall and about 9” wide. With one of us packing bricks and the other incorporating the bricks in to the wall, the work was MUCH faster and the walls were far more even.


Jen applies fire cob plaster to the interior surfaces. The smoke shelf is clearly visible in this photo.


Cob bricks ready to be incorporated in to the fireplace wall.


When the cob had locked up sufficiently we dropped the wooden arch form (it was built so that the legs could be removed, thus allowing the arch to literally fall). We found that as we built we had problems with slumping and bulging, particularly on the front wall. We used a wood saw to shave down the front wall to even it out and in doing so accidentally created an arched relief above the front opening. We ended up really liking it so we kept it like that. Once the fireplace was built up to a height of 2’ above the smoke shelf we began tapering it in to form the chimney. We used the wood saw to shape the outside in to shoulders that arced in to the chimney. Our chimney has a rectangular footprint with soft inner corners, 21”x10”. It was around this time that we started to believe that this project might actually work! Up until this point we had been very worried it would end in an embarrassing failure. It is really amazing the magic that can be worked on cob with a wood saw and a bit of patience.


FP 20_starwatermark
The final shape is starting to appear.


The chimney grew until it seemed he right height, stopping periodically to plaster the interior in fire cob. We often had to trim off sticky-outty bits of straw on the interior to get the plaster to stick well.


FP 21_starwatermark
Jen works on the chimney.


FP 22_starwatermark


The outside was still pretty wonky so we tidied it up with a lot of blobbing on cob to the parts that needed blobbing up, and shaving down the parts that looked overly blobby. We spent A LOT of time doing this, working it over several days to observe it in different light, from different angles, and asking everyone who would take the time to stop and look whether any parts seemed out of character. It made a huge difference! Because we had initially built the side walls to exactly follow the shape of the firebox, our fireplace ended up with strange backward angled side walls. We went back and built these up to make the fireplace sidewalls more square with the front/back. We also started building up the benches. We got tired of how much cob it was taking so we cheated a bit and built up the interior of the benches with a fair number of rocks. Boy was it worth it. We also decided to cut little decorative alcoves out of the side walls of the fireplace for resting a drink or nesting a candle. The alcove bottoms were fitted with cedar slabs to create a wooden base to them.


FP 23_starwatermark
Alcove and bench under construction. Roof posts going in.


The first fire was a huge success! In this photo you can see the wet, and dried smooth, cob where we have been working to tidy up the shape of the fireplace. We ended up extending the left side bench to form a T intersection at the end of it, a great suggestion made by Jen’s mom.


FP 24_starwatermark


We always intended to lime plaster our fireplace but decided it would benefit from a roof structure to keep the driving Vancouver rain from pounding it down. We built ours using a four post system and an A frame slatted in galvanized corrugated roofing. We worked out the dimensions of it using Google Sketch Up.


FP 24a_starwatermark
Sketchup roof model.


FP 25_starwatermark
Building the roof.


We also found an old fireplace screen in an alley and hauled it back home to protect ourselves from the flying sparks.


FP 26_starwatermark
Salvaged fireplace screen.


Ryan did an incredible job of creating the fishbone slat work!


FP 27_starwatermark


FP 28_starwatermark


In the peak of the roof Jen cut barn swallow silhouettes out of metal. We chose swallows for two reasons. First, for over 60 years barn swallows have raised their babies annually in our barn, making them a cherished fixture for our family. Secondly, they build their nests from little balls of clay mud mixed with straw and horse hair, just like our fireplace!


FP 28a_starwatermark


We made mason jar candle lanterns which we hung from the roof and put in the alcoves. The fireplace really came to life!




It was now time to plaster. WEAR GLOVES (and ideally goggles) WHEN WORKING WITH LIME. It is terribly drying and can be very caustic to the eyes and skin.

Our plan was to attempt to plaster the entire fireplace in a tadelakt finish. We first applied a scratch coat plaster, made from 1 part NHL2 lime combined with 5 parts sharp coarse sand and a liberal amount of straw. The straw, as we understand it, strengthens the plaster, acting like little pieces of rebar to prevent cracking. The fireplace was moistened with a hose and the plaster applied evenly but roughly so as to create a coarse surface on to which the finishing plaster could key in. Lime plaster is incredible but requires a thorough understanding of its chemistry in order to get a strong finished product. We highly recommend that you read up about it (references below) prior to working with it the first time. If lime plaster is allowed to dry too quickly it will be weak, crack, and slough.


FP 29_starwatermark


FP 30_starwatermark
The t-bench, wooden base in alcove, roof finished.


To maximize the time that our plaster remained damp, we covered the entire fireplace in wet towels which we kept damp by periodically misting them with the hose. We left our fireplace covered for a week. It looked ridiculous but worked well!


FP 31_starwatermark


We found that even after the scratch coat was applied that we still had a lot of low areas on our walls. Prior to applying the finishing coat, we applied a straightening coat of plaster, made from 1 part NHL2 combined with 4 parts sand sifted through window screen and short pieces of straw (we chopped up the straw by going at it with a weed whacker in a garbage pail). This coat minimized the amount of finishing plaster we needed to use (far more expensive). It would have been smarter to do a clay based straightening coat prior to the lime scratch coat to save on lime costs. We’ll know that for next time.

It was finally time to apply the finishing coat. Prior to starting this final coat we read extensively about tadelakt, a highly water resistant, silky smooth plaster technique originally from Morocco. Moroccan lime is apparently burned fairly inefficiently, resulting in a portion of the lime left in its unburned, limestone state. Inefficiently burned lime therefore has its own built in aggregate. North American lime is burned to near completion and is as such largely devoid of aggregate (hence the need to add sand when working with it). Some believe that tadelakt can only be produced using Moroccan lime. We disagree.

Some people use marble dust (limestone) as the aggregate when using North American lime to make tadelakt. This most closely reproduces the lime found in Morocco. The reality however is that Moroccan plasterers made tadelakt using Moroccan lime because that is what they happened to have on hand. Had they only had access to purely burnt lime and west coast sand, they would have found other ways to create the same finish. The finish of tadelakt, we have found, is contingent far more on the way in which the plaster is worked, rather than its elemental composition. We still have a lot to learn about the actual application of it and have many more projects lined up on which to practice our technique.

We considered burning our own limestone to get a lime naturally rich in aggregate but decided that was too much trouble. We then considered using marble dust as the aggregate but it was expensive and we had boat loads of sand that needed using up. In the end we decided to try using plain old sand, sifted through fine window screen and combined 1 part NHL2 to 2 parts sand. It worked great.

Tadelakt by definition has a network of little cracks throughout which apparently look beautiful and do nothing to undermine the function of the plaster. Being terrified of cracks in plaster however we used a 2:1 ratio and added a good amount of cleaned and carded horse fur (coat, not mane/tail hair), collected during our horses’ spring shed, to our plaster. We found that if we eliminated the horse hair and used a 1:1 ratio, the plaster finish cracked in the same way as the photos we have seen of Moroccan tadelakt. Maybe we will be brave enough to try this for our next project.

We also coloured the plaster with tint purchased online from For our fireplace we used nearly 3kg of Rose Earth and 1kg of Colonial Raw Sienna colour but have learned through experience that the amount of tint needed is highly variable and virtually impossible to predict. We have found Earthpigments to be an incredible source of pigment and have been exceedingly happy with the service and products we have received.

Despite the fact that NHL2 is a hydraulic lime, it is so weakly hydraulic that in many ways it behaves like a non-hydraulic lime, or lime putty. With this in mind, we found our plaster MUCH more enjoyable to work if we prepared it and then left it to slake overnight. When we knocked it up the next morning it was buttery and silky, an absolute dream to work with. If the plaster was kept moist with daily knocking up, we were able to store it for weeks and it only got better over time.

Prior to attempting to tadelakt the fireplace, we made a couple of bird baths. It was well worth the practice!


Tadelakt bird baths holding water.


Time to tadelakt the fireplace. The fireplace was again moistened with the hose on a mist setting. Then we began plastering, first applying the plaster with a putty knife and then working it smooth with the knife or a float like icing a giant cake. As we worked, areas that had been previously plastered began to lock up. We regularly misted the fresh plaster with water from a spray bottle to keep it from drying and cracking. Do not plaster in the direct sun or you will never be able to keep up with keeping the plaster moist! We rigged a ridiculous network of tarps that acted as sun shades.

As the plaster locks up it transitions through a phase that is much like plasticine. It is malleable but resists being dented by finger pressure. Once it was at this stage we reworked it, polishing in firm tight circles first with the lid of a yogurt container (with the rim cut off) and then with ultra smooth river stones. As we plastered our fireplace we were continually going back to areas already plastered to polish as well as frequently misting with water from the spray bottle.

The length of time needed to let the plaster set is highly variable and depends on the outside temperature, the humidity, and the suction force of the material on to which it is being applied and probably a whole host of others things that we aren’t even aware of. We spent a lot of time practicing working with plaster, first on small bricks of cob, and then on little birdbaths that we made in order to learn the feel of it. No matter how many books we read, we didn’t understand how to work it until we started practicing. It also takes time to learn the feel of working it. Be prepared for immense frustration and to have to make many attempts before feeling confident.

Interestingly, the more pressure that is applied when polishing, the deeper the final colour that is achieved. Knowing when to stop polishing is also learned by feel. Further, if the plaster is left to set up too long it becomes too hard to work and can’t be polished smooth. It took us two full (>12h) days of plastering to complete the finishing coat, with 2 of us working on it.

In the photo below you can see the contrast between the plaster that has been polished and the plaster that has yet to be polished. Jen is applying fresh plaster around the horseshoes, she has already polished the plaster above this. The shoes are attached to the fireplace over top of the straightening coat using long nails driven through the plaster in to the cob following pilot holes created by a drill.


FP 32_starwatermark


Using the sand in the ratio outlined above, we were able to get a silky smooth and glistening finish with polishing. We were astounded. Having never touched Moroccan plaster we are not in a position to compare the final outcomes however we are absolutely thrilled with the finish we achieved and have a hard time imagining how it could be any smoother. Also, as we polished we found that the sand crystals began to glisten, creating this magical sparkling effect when the sun hits the fireplace. The horse hair completely disappeared in to the plaster and we can’t tell that it’s there except when we compare it to areas devoid of hair which have notable tiny cracks in the finish.

We again covered this plaster in damp towels which we misted periodically to allow for a slow set.

We then applied the soap finish. As we understand it, applying soap to lime plaster creates a saponification reaction that affords the plaster remarkable water resistance while maintaining breathability. We made our own olive oil and lye soap. Once this had cured we blended up about a tbsp of shaved soap with roughly 2 cups of water in the food processor to create a thick, sudsy foam. Face cloths and dishrags were wrapped in plastic bags to create a little pillow. The pillow was then dipped in the suds and buffed over the lime. Be careful that the bags don’t have any print on them or it will rub off (we discovered) on to the plaster. The plaster became immediately smoother, deeper in colour, shinier, and water resistant.


FP 33_starwatermark
Soaping the tadelakt.


In the end we applied 5 coats of soap mixture over several days. What’s amazing is that you can wash the plaster periodically with the suds mixture which both cleans and conditions it! Incredible!

We are absolutely thrilled with our fireplace and have had countless evening fires, even in to the winter. The roof does a good job of protecting the fireplace from driving rain and the plaster so far seems more than capable of tolerating the wetness that does reach it.


Fire in a rare Vancouver snow.


While our plaster is marble smooth, shiny, and water resistant it was made using locally available materials and was deliberately crafted to prevent cracks. Traditional tadelakt is made from lime and soap unique to Morocco and is crafted to deliberately promote artistic micro cracks. As such, we hesitate to call our fireplace a traditional tadelakt but are hard pressed to see any real difference between ours and what is found in Morocco as far as function and texture with the obvious exception of ours being devoid of cracks. We believe that we have created a North American West Coast tadelakt that we are very proud of.


FP 34_starwatermark


FP 35_starwatermark


FP 37_starwatermark


FP 36_starwatermark


FP 38_starwatermark


  1. Building with Cob: A Step by Step Guide. Adam Weismann & Katy Bryce.
  2. Using Natural Finishes: Lime and Earth Based Plasters, Renders & Paints. Adam Weismann & Katy Bryce.
  3. Tadelakt. Michael Johannes Ochs.

11 Replies to “Building the Cob and Tadelakt Outdoor Fireplace – Homemade Tadelakt!”

  1. Hi Jen, my name is Ana saw your blog on Cob & Tadelakt and was wondering if you bought your black soap or just made it? If you made it could you post how to do it. Beautiful work btw.

  2. Hi Ana

    Thanks for the comment!

    Sorry for the delay, I am just getting back from out of town.
    I made my own soap. From what I understand, the specific type of soap is largely unimportant. The tadelakt finish is a chemical reaction between the calcium carbonate in the lime and the glycerin back backbone in the soap, essentially exchanging the carbonate for fatty acid salts. This is a saponification reaction and is what gives tadelakt its water repellency and its velvety shine.

    When practicing I used commercially available hard soaps including Dove and Ivory, both of which worked well.

    For my fireplace and tandoor I used an olive oil soap that I made at home based on the recipe available here: There are different ratios of oil to lye that will yield different outcomes so if you are inclined, you might want to try a few different recipes.

    I think you could use any soap recipe that yields a hard oil based soap for tadelakt as I did not perceive any difference between the different runs. I only made my own soap for the fun of it and to make me feel that I had truly made a home made tadelakt in the most essential of ways.

    Morocco uses what they have on hand. You should likewise use what you can get your hands on easily in your corner of the world.

    I hope this helps.


  3. Hi Jen,

    Thanks a bundle ^^ for a detailed & complete answer 😀
    I’ve been researching & researching looking for recipe’s of the black soap. Seems like I was searching in the wrong places. I’ve attempted in making Beldi(BlackSoap) first batch turned out to be Castile Soap wasn’t familiar there was 2 types of lye. xD Second batch was a flop when I used KOH.

    Thank you greatly for the link. And the assurance. I forgot about the post I sent out. Funny I’ve been up all night watching soap making videos and joined a soap forum just to make black soap & I have exhausted myself since it’s 7:13am & I’m about to pass out. But after reading your reply like a surge of energy came back to me xD doesn’t really make sense. But I’m going out right now and try the ivory soap you suggested 😀


  4. Glad to help. Let me know how your tadelakt experience turns out. I’m always curious to learn what others discover as they go about their projects. Also don’t hesitate to let me know if you have any further questions 🙂


  5. Your instructions and illustrations are wonderful, and the best source of information on tadelakt application that I can find. I agree that this is a very frustrating process until you understand and practice. I am wondering if I can apply the soap suds with a soft bristle paint brush….Also, I feel confident I can use my homemade soap, thanks to your comment!

    1. Hi Kathy!

      Thank you for the kind comment! I have moved it over to this domain from our old, obsolete blogger domain.
      I’m glad you found our blog helpful and I hope you are getting lots of tadelakting done. You can absolutely use your own hand made soap. I’m sure you can apply it with a paint brush but you’ll want to polish in it. You’ll notice it change dramatically when the soap in worked in to the plaster. I’d recommend using a plastic bag pillow as described in the post. It works really well for us.


Leave a Reply

Your email address will not be published. Required fields are marked *