Last week I got to go out and help out with the earth shelter being constructed near Battle Creek.  As usual, there is nothing as enlightening as seeing things first hand…  If you are planning on building an earth sheltered home, I really recommend you find one in progress and visit as often as they will let you ;^)

Progress

When I got there, I saw that Scott (the home owner, GC and whatever other role is needed) had done a lot of work since the last visit in September.  He had completed the rebar work, including a second layer on the larger dome.  He had setup the ICF (Insulated Concrete Forms) on the south side and a pumper truck with a tall boom had come out to fill the forms for that front wall.  His carpenter had also installed most of the windows.

While other jobs were being done by professionals, the home owner had started on wiring up the polystyrene on the inside of the domes.  The polystyrene is used as backing to prevent the shotcrete (sprayed concrete) from just blowing right thru.  The basic process, as specified by Formworks Inc, involves carefully cutting sheets  of 1 inch thick polystyrene (what white coffee cups are made of) to fit in the space between the IBeams.  Two overlapping layers of butt jointed polystyrene need to be placed at once.  The attachment is made, starting on the outside of the dome, by taking a “u” shaped piece of wire placed around an intersection in the rebar and pushing it thru two layers of polystyrene.   The wire is wrapped around a thin wood baton that prevents it from simply pulling thru the polystyrene.  The polystyrene is installed loosely to the rebar, with at least “one finger” of slack in the wire, to allow the shotcrete to envelope the rebar.  Cutting the sheets to fit the 6 ft spans between the IBeams is a bit of a hassle, but making the precise cuts to fit two overlapping offset layers and wiring it all up is very time consuming.  Working high off the ground, where the shapes only get trickier, adds its own special pain.    Seeing that the cold weather was coming and he needed to speed things up, the homeowner called in some hired hands so he could be ready to shotcrete before the really cold weather hit.  They got the job finished in 10 working days.

This is a wider view of the insulation inside the smaller garage vault. A lot of effort went into this part of the puzzle.  You can see the light coming thru both layers.  I suspect things would have gone more smoothly if the outer cracks had been taped to prevent shotcrete from getting between the layers, but that is probably easier said than done with all the rebar in the way.

The “Formworks” process starts with the polystyrene on the inside as a backing for the shotcrete, but it doesn’t stay there.  After shotcrete, the polystyrene is removed and then used again on the outside as insulation (out-sulation) and protection for the waterproofing.  Since the insulation is usually pretty cut down or broken up by then, it doesn’t really form a nice continuous layer and additional insulation will probably be needed…

Electrical boxes and conduit were also set in place between the polystyrene and the rebar.  These will eventually be enveloped by the shotcrete, but open to the inside.  Of course, the electrical inspection had to be done before they “closed up the walls.”

 

 

Shotcrete

This was the shotcrete crew’s first earth sheltered home, but they have lots experience with various other shotcrete structures, including some with shotcrete placed overhead.   I first spoke to Nate well over a year ago.  He stood out from the shotcrete guys I spoke to as someone who not only knew what earth sheltering was, but thought it was a good idea.  (one of the others said “you mean like a zoo habitat?  Why would you want to do that?”)  I meet up with him in a McDonald’s and he had almost as many pictures of earth sheltered homes on his computer as I had on mine… It turned out that he had already had meeting with another potential earth sheltered home builder who was also planning an earth sheltered house.  Not sure if Scott had to convince or educate him about earth sheltering, but I appreciated that he was already interested.  One thing lead to another, and I got Nate to introduce me to Scott so I could get some first hand experience.  Plus, how often do you get to see a contractor work on a similar project before you hire him?

Shotcrete uses a cement pump and a powerful air-compressor to blast (with air pressure) low slump cement onto the structure.  This concrete mix has much less water than a pourable mix, which results in higher strength (6500 psi in a week, 8000 psi in 28 days).  Also, the heavy sprayed cement is moving quickly (inertia) and compacts tightly as it hits the rebar and previous shotcrete without needing any vibration.  The result is a dense and very strong cement structure with no seams or cold joints.  One of the main benefits of shotcrete is that you don’t need traditional formwork, which means you are not as cost-constrained to building with straight walls.  However, it does help to have some sort of backing to control the shape and prevent some of the cement from being wasted (by just shooting right thru the wall).   The shotcrete stiffens quickly and locks on to the rebar and previous cement.  It easily spans over the many small (1 or 2 inch) gaps in the insulation.  The “gunner” starts with a thin layer and keeps the gun moving so the shotcrete has time to set before too much is added (so it doesn’t just slump off the wall).  They keep moving and gradually adding thickness to the cement until it reaches the specifications.  In this case, the engineered drawings specified 12 inches at the base tapering down to about 8 inches at the 10 ft level and then down to 4.5 inches at the top of the larger dome.  The top of the smaller dome only specified 4 inches.

With shotcrete design, some curvature is actually an asset as it helps the wall stand on its own without as much bracing (A curved piece of paper can stand on its edge) and a convex curve resists earth loads with less thickness or reinforcement…

 

I am sure the Shotcrete guys had lots of practice/experience with swimming pools, but this slippery, flimsy and loosely-wired polystyrene backing was new and there was a learning curve.  While the shotcrete thickness can be applied over several days without a “cold joint“,  they generally want to apply each pass as thickly as they can while moving around…  However, while that shotcrete is still setting, its weight, along with the impact force, is all against the polystyrene.  Also, since the polystyrene is only loosely wired to the rebar, it moved around alot as the shotcrete hit it and made a lot of noise (until some shotcrete thickeness built up).  During this shifting, gaps opened up, etc.   It is actually amazing that the easily broken stuff didn’t just tear off…  There were a few bulges where the rebar sagged and actually pushed in (from convex to concave, oil canning)….

There were also a few blowouts, where the polystyrene did actually fall apart.   It appeared that most of the blowouts happened when the weight of the wet shotcrete pushed the rebar so far inward that the polystyrene was stretched to the breaking point.   The other failure mode was when shotcrete got in between the layers of polystyrene and was able to flex and break of a piece of the inner layer…  I think that taping the cracks on the outside could have prevented this, but with all the rebar in the way, that is probably easier said than done.   When there was a blowout, the polystyrene and hundreds of pounds of cement came down with an awful crashing sound (I was inside about 25 ft away at the time of the largest one).

Before each blowout, we saw shotcrete pea stone pouring in between the overlapping layers and then the wall started to bulge inward.  If you can stop it then, you can prevent the blowout!  Then the shotcrete between the layers started to push into the vault and cracked off corners of the inner layer.  These cracks freed the batons and then it all came down.  All in just seconds…   Since shotcrete between the layers started the chain of events, closing those outer gaps (with tape) is probably the most important preventative measure.  Taping the inner cracks is not nearly as important.

Scott handled the roof caving in pretty calmly.  While the shotcrete gunners proceeded (with more caution) on to other areas, Scott called in a friend and we quickly assembled some scaffolding and replaced the insulation.  We also added more batons to stiffen similar areas, but I am not sure how much difference that made since it wasn’t the batons or polystyrene that failed.

I was really worried about further blowouts, particularly since we hadn’t got to the even larger home dome yet.  But it turned out that “practicing” on the garage was a good idea.  I was only there for day 1, but I was told that they didn’t have any blowouts or issues on the main vault over the next few days.

 

I had posted a time-lapse video of the process right here, but after 5 years, the shotcrete contractor asked us to take it down and remove his name from these posts because he was getting criticism in the comment section. I am pretty sure he had no legal grounds, but I decided to oblige him anyway.

 

Lessons Learned

Between scrambling to shore up bulges or fix blowouts, there was time to chat with the homeowners, shotcrete crew and cement mixer drivers on the site.   We talked about a range of things from the cost of job site insurance to the hidden costs of heating systems to the suggestion to install cheap steel doors to close up the house (security) during construction (while the nice doors are safely away from contractors dents and scrapes.)   Some shovel based civil engineering had been done to clear water from the site, but it was clear that the shotcrete crew was struggling a little with the steep banks close to the site (although sometimes it seemed helpful to shoot from the banks.)  I reinforced my mental note to grade around my site more carefully.

During my previous visit, I was impressed with how well the Formworks steel and rebar system was designed, but this time, I was very glad I was not following their process for the polystyrene shotcrete backing.   The polystyrene worked, but it seemed like my pegboard/particleboard plan would be better in almost every respect.  The polystyrene backing required a lot of time-consuming cutting to get the two overlapping layers.  This is easier when you have the IBeam flange to hide the edge or when your arches are parallel, but I would have neither.  Also, with the Formworks plan, the polystyrene is inside the IBeams and tied, somewhat loosely, to the rebar to provide room for the concrete.   This loose polystyrene shifts a lot when the rebar hits it, and in some cases, allowed shotcrete between the layers.  In all the “post shotcrete” pics I have seen on the Formworks website, this process leaves a somewhat rough final finish on the inside of the dome because the edges never quite line up and the gaps between butt joints are at least an inch deep.  Conversely, the polystyrene itself is actually too smooth, which makes it difficult for the shotcrete to grip.  It is also brittle (easily broken).  The polystyrene is also relatively expensive and I don’t buy the argument that the cost is offset because it can be “re-used” as insulation.  By the time you take that polystyrene down, it is so chopped up that it will be impossible to form any sort of continuous layer, even overlapping it like shingles.

I have seen burlap used on other earth shelters (such as the Project Michigan Earth Shelter videos available on Vimeo).  It is cheaper and doesn’t need to be removed, but it sags and gives a very ugly appearance on the inside that is not something you can just plaster over…

I had preferred the pegboard plan.  It is cheaper than the polystyrene, and strong enough that one layer is sufficient.  It is harder to cut, but if you overlap the pieces a little and not need to cut as often.  Also, with the 1/2 inch tube arches, you can tie it tightly to the inside of the arches and properly control the cement thickness around the rebar.  Also, the pegboard provides better grip for the shotcrete and would prevent some sagging.   It also provides better control of the final shape and the little quarter inch thick posts left over after the pegboard is removed provide a nice final surface for plastering the inside of the vaults.  We actually plan to use “SpecFinish”, which is a fine sand-based shotcrete, on the inside of the vaults.

However, one of the shotcrete guys pointed out that the downside of the pegboard was its susceptibility to moisture.  While the polymerized linseed oil on the surface of the hardboard gives it some water resistance, the drilled holes provide easy access for water to soak the wood fibers inside.  I would need to worry about rainy weather and the moisture of the shotcrete its self.   Also, once shotcrete mushroomed thru the holes, it may have been much harder than the polystyrene to remove later.

Apparently, the shotcrete crew had recently shot a movie set where expanded metal lath was used as the backing  (some storm-related movie, “Black Clouds” or something like that where they needed to do a flood scene).  They said the metal lath was the easiest thing they ever worked with.  The Metal lath has all the advantages of the peg board, except it is a little more expensive (20%) at roughly 35 cents per square ft.  It also has some additional advantages such as; being relatively impervious to weather, adding reinforcement and not needing to be removed after the shotcrete is applied.  Metal lath can be cut-to-fit on site with a hand-held grinder, or the sheets can just be overlapped and wired together.  The metal lath is stronger than the other options and can actually be walked on during the shoot (but mine will be below the rebar anyway).  It also holds its shape under the weight and impact of the shotcrete better than most other backings.   The shotcrete comes thru the metal lath just enough to mushroom out the other side and hide most of the metal.  This provides a nice evenly-rough surface on the inside, ideal for finishing.  I already talked to my architect and engineer about switching to metal lath.

I also noticed how much pea stone was “rebounding” off the wall and being raked away…  In my design, vaults meet at the bottom and I was concerned about where all that pea stone getting stuck between the vaults.  I discussed this with the shotcrete contractor and he said it would be a bit of a problem.  We will need to remove it, even if it means scooping it out with buckets…  On the other hand, my vaults are not nearly as tall as the earth shelter near battle creek, so it won’t be as much pea stone anyway.

 

Next

Next, I hope to head out to help with the waterproofing on the shelter.  More on that later.