Virtual Build
Over the past couple weeks, I put in half an hour here and there in the evenings and moved the virtual build slowly along. Of course, I would be thrilled to get this virtual build speed in the real world ;^) Initially, the plan was just to illustrate the building process for various subs that would be helping. I planned to just show the construction of the basement, central tower and how the ribs would be setup. However, as you may have seen over the past couple posts, the virtual build has already helped find and solve so many problems that I have decided to press on with it. I added the garage and I am currently working on the bedroom wing. Next, I will work on the front of the house.
Virtual build as of Feb2nd, the garage, mezz and other areas are mostly complete, I am working on the bedroom wing and will work on the front of the house later. You can already see that this house design is very original.
I had been thinking a lot about how to lay out the ribs while I was casting them, I thought it might be a good idea to illustrate that with the virtual build. For a long time, our plan has been to pour the garage slab early in the build, but only setup the rear of the quonset hut. I could use the partial construction as a covered workshop in the back while I setup the rib forms on the front half of the slab. Without the roof in the way, I could use a crate to lift the 4500lb concrete ribs up and over to where they would be set.
I had planned to make a left and a right rib form because it would save me needing to flip and polish the “back” side of one that would be exposed against an end wall. I had already worked out (mathematically) that I could fit two of these on the slab in front of the partial quonset.
Now I was working on my gantt chart and considering how much time would be taken making these only two at a time. I also considered how much extra cost would be incurred with the multiple crane visits (to set them up 2 at a time). We had already planned to put a slab in-front of the garage (even if the rest of the driveway will be gravel), so the boys could play basketball. I started considering the option of placing the slab earlier in the process and building a few more forms. Even if the extra forms cost an additional 600$ each, I could make 2 more and save on 2 expensive crane visits and a couple weeks of time. I would probably at least break even on money, but save time.
There would be no additional cost to adding the concrete pad early since it was already planned for later in the process, but it may later be seen as “in the way” as construction equipment would need to be careful not to crack it…
So, I added the ribs to my garage model… You can see that it would be very difficult to fit a 3rd rib on the pad. Yes I tried other configurations, but I need room for the forms around the ribs and space too work, and I couldn’t let them go under the quonset or the crane wouldn’t be able to lift them up without dragging them. However, for the other ribs, I decided to make two left hand ribs and they can be put closer together…
Then it occurred to me that I needed to go back and reconsider my earlier decision about wanting to start with a left and a right rib to save time/money polishing (because one side would be placed against a wall.) I decided that I would save more money and hassle with the more compact arrangement of same-side ribs. In the software, it is pretty trivial to make the necessary changes and voila! Four ribs in a compact arrangement on the one slab.
I will update the Gantt to show building two forms first, then working on the second two forms while the first two ribs cure… Then I could position all 10 ribs in only 3 crane visits instead of 5. Of course, I still need to polish the back sides of the ribs, so I will need to flip them over before I set them, but I can place them in the (soft) dirt for that and I think I can get that done with only one extra crane visit for the full set. It seems like a plan for now.
Other revelations included that the steel stud layout for the bedroom was messed up by the architect (yup, I checked my original notes). It was a classic symptom of 2D design where the various views, created separately, were not actually compatible or build-able. I went back to check my original notes and sketches that I had sent in and they were correct, so I guess the architect just didn’t understand. I will document that another time. Lets just say my errata list is growing.
Sourcing
I have started hearing back from various contractors and the bids are looking much better. I got a very reasonable excavation bid to go with my good footings bid. I got bids on hooking up my electricity (a very reasonable $285 to setup the temp construction meter and then about $4.60/ft to run the permanent cable and setup for 400 amp service. I am waiting on an electrical and HVAC update that should be in early next week. It is clear that I still need to find a reasonable plumber bid (the 6 I got last year were all either too high or too un-reliable).
I also got prices on HDPE pipe for various diameters… Trying to understand the pricing structure better, I divided the price by area, weight, etc. and quite reasonably, it turns out to be priced by weight. $1.25/lb, delivered. For 8 inch pipe with a 1/4 inch wall (HDPE 8″ DR 32.5), that comes to about $3.83/ft. I need about 450ft for an earth tube loop, plus about 350ft for internal duct work in the house… So that goes into the estimates.
Shotcrete meet
A highlight of last week (for me anyway; I am not sure how your week went ;^) was a meet with my most-likely shotcrete guy. I have been talking to a couple other potential shotcrete contractors, but one is just not big enough scale and the other is not really sure about the whole earth sheltered concept. At this point, I trust Nate more than the others and I think his prices seem fair. The biggest problem is that he is hard to get ahold of. It has been more than a year since we managed a meeting. My project is just too unusual for him to quote confidently, so he had agreed to a time and materials quote, but I needed a better idea of how long he thought things would take and I still hadn’t got a quote on some aspects of the build, such as the specfinish on the inside.
I drove out to his place, which is about an hour from my current home and about 40 minutes from the building site (its a big triangle around Ann Arbor, MI). He had a couple big friendly dogs and a nice sized kitchen table to lay the plans out. I set up my computer and showed him my virtual build, my Gantt chart, etc. My main goals were to make sure that the plan for the build made sense to a professional and experienced shotcrete guy and to get good numbers for budgeting purposes. I showed him the virtual model and we talked about practical things like how to get the rebound (shotcrete that doesn’t stick to the wall and is, therefore landfill) out of the basement. We talked about the possibility that some of the non-load bearing walls could be built hollow (or filled with insulation) and still covered with a thin coat of shotcrete to match the more solid walls. We talked about hiring some of his guys to help tie the rebar (along with other tasks), and how quick they could work. He even shared some trade secrets about the fastest ways to tie rebar, plaster walls, etc.
Along the way thru the virtual model, we kept referring back to the Gantt chart, which included things like the dates for each of the 4 shotcrete phases and the amounts of shotcrete that needed to be applied in each phase. He factored in if it was high work or regular walls (which he kept referring to as “money walls”), etc. We discussed if the dates were good for him. He was a bit concerned about the first shotcrete date because it was in May, which is prime swimming pool season. The other dates are past swimming pool season, so he will be glad to keep his schedule busy.
We discussed equipment that I would need to rent and what his crew would bring (and the associated costs). No surprises there except that he mentioned he would bring lots of scaffolding at no additional cost, he just wanted me to make sure the floor was level enough to move them around.
We discussed how much notice I would need to give to schedule his crew.
The main rough patch came when I got to the part about using his gunite machine to spray the specfinish along the underside of the vaults. Nate did not want to do that. Apparently, it is messy horrible work. See this pic from monolithic.org where they create inflated fabric domes and then coat the insides with shotcrete to form the structure. Note the full body coverage including saran-wrap on the helmet. When the operators face gets covered, he can pull on the roll to get a clean section. My plan had most of the shotcrete structure applied from the outside and only a very thin (3/8″) layer applied on the inside, but still…
Nate said he would rather just have his guys apply the plaster by hand, but that sounds rough and slow to me. We talked about other alternatives including spaying on the ceiling with a drywall hopper gun or having an acoustic ceiling company come in and take care of it. I had already got a quote on acoustical ceilings, but my wife didn’t like the samples I brought back and didn’t want to consider it at the time.
Another option may be that I would use the same glue up styrofoam ceiling tiles that I planned to use to form (impress) the ribs. It may look good to have the same pattern in the vaults between the ribs, but I am a bit concerned about fitting the square pattern to the curved and radiating vault shapes and it wouldn’t work at all in the compound curved bedroom vaults.
I left the meeting with a few notes on minor changes that I needed to make to the Gantt and process and some homework to Google search a number of things that were discussed (such as a rebar tie belt with a reel). I also have the costs I will need to complete this portion of the budget.
Virtual Build
Last post, we talked about finding problems with the architects drawings. Mostly, the issue was just that the drawing was not quite to scale. The dimensions shown were not actually the dimensions used. I had followed the dimensions as shown when building my virtual model, so by the time I got around to fitting the kitchen door, it didn’t work. Some things I could fix, such as the radius of certain rooms that had been mislabeled by a number of inches. Other things, such as the width of the arch shown on the kitchen wall elevation, were just plain wrong and I had to figure out how to deal with the misfit.
Note: In an earlier post and on a page, I talked about how I went with the architect who offered only 2D drawings because it saved me a lot of money on this difficult-to-3D-render home. At the time I understood that 3D is better for most rectilinear homes because it does help to find these problems, changes are properly propagated, etc. But for this complicated unconventional home, I reasoned that I would be paying a lot for the 3D cad skills when I really only needed the architects design skills and the final builders would only need 2D drawings. I never did get a lot of architectural input on my 2D design, and now that I am doing the 3D model myself, I am finding all the problems that a 3D architect would have found. Choose wisely, even if I am not really sure what the wise choice would have been yet.
I am using SpaceClaim to build my virtual model. It doesn’t have a lot of the fancy textures or architectural features like widets for making sloped roofs or easily adding doors to walls. I also have Autodesk Revit, which does have those features. I used it for my earlier models (pre architect) and that tool was great for layout, etc, but I found it much harder for the complex geometry of my roof. SpaceClaim can handle the complex geometry. Spaceclaim is also great and modifying a geometry to fix a problem. It is a “direct modeler”, so you can grab any surface and “Pull” it to adjust it. Everything gets taken care of along the way. This made fixing the model pretty easy.
I needed to reduce the radius by 4 inches. Even on complex objects like this concrete over my quad-deck, it was easy to pull things into place…
Other tools, like Rhino3D, are popular with architects and could handle the geometry and has much better rending functionality (and my architect’s junior guy did model some aspects of my design in Rhino3D), but it doesn’t have the tools for easy modification and I didn’t have a licensed copy. I am told that Sketchup Pro, very popular with architects, could handle this. However, I tried the very popular free version, aka “Sketchup-Make” and it could not handle the ends of the vaults in the circular portion of the house (although it would be great for something more rectolinear).
Once the radius problems were fixed (and I added all the changes to my errata sheet), the door still didn’t fit because of the out of scale arch used in the architects elevation. I had to think about my options. I was not willing to re-scale the rib to match the architects mistake in one drawing. I decided that the rib at that location was “architectural”, not structural. This meant that I didn’t want to move it outward or upward and lose its architectural look/fit with the rest of the space. It also meant that I could cut the spandrel without needing to re-engineer the arch because that arch was not really bearing the load of the roof like the other arches were.
This left me with two main options.
1) I could cut the arch to fit the door. This would give me a full rectangle door to work with. I could get my “architectural” look back somewhat by coloring the door some how to continue the arch. Maybe I would add a veneer of granite, or stained glass or just stain or paint, shaped so that when the door was closed the concrete arch lines were continuous across it. Well, my wife did not like that idea. She is concerned about the structural aspects and she is probably concerned that it will be more work (the house is enough work as it is and she hates it when I add to my potential work load).
2) I could accept the fact that the arch crosses a big corner of the door. To test this idea, I have actually put duct tape across the corner of my office door in a way that matches the profile of the arch crossing the kitchen door. It has been there a while now and I have not minded it at all. Even if I brush my shoulders against the door frame, my head still does not hit the duct tape corner. If I go with this plan, I have multiple ways of proceeding…
2a) I could cut out a section (or just prevent concrete from forming in that middle space) so that I could fit a rectangular sliding door frame tucked into the arch. If I take out 4 inches of concrete from the middle of the spandrel, there will still be 4 inches on either side. Plenty of structure for an arch that is filled in on the underside with concrete anyway. The door would slide into the wall between the kitchen and basement stairs. I like this idea, but it will take some careful planning if I am to form the concrete rib with the void in exactly the right place… (planning the wall void is easier because I am building it right in place over a framework.)
2b) I could simply hang a sliding barn door (but a modern looking one with nice or hidden hardware) on the outside of this kitchen doorway. The door would be hanging in the mudroom and could slide the opposite way along the mudroom wall, so I wouldn’t need any voids in the wall between the kitchen and basement stairs. This is Sherri’s preference, at least partially because she thinks it will be easier to implement. I think the architect may also have suggested it at one point (because he didn’t know how I would get the mechanism inside the concrete wall). I don’t like the “fit” of it as much, but I will try to keep an open mind and think about it some more.
In the mean time, here is how things are looking in the kitchen (I modeled in some cabinets to make sure it all fit)… The three open blocks above the cupboards will be 8″ glass blocks and are there to let light from the main living space into the basement stair well.
And here is a wider view of the north side of the house (the original plan was to virtual-build just the section over the basement). Of course, this is just the initial concrete structure (plus door bucks). No earth cover, windows, etc. You can see that the mezzanine windows have been moved closer together to allow the dirt to cover the roof better. Inside, I added other details, including the spiral stairs, etc. Maybe I will include some of those pics in the gallery at the bottom.
Fusion welding HDPE Plastic Pipe
I got a section (about 6 ft) of 8 inch HDPE pipe from a contractor a while ago. It was old and cruddy and maybe had a bit of oil on it, but I took it so I could experiment with it. The expensive part about building earth tubes with HDPE plastic is that you have to hire someone to fusion weld them together… Or at least, you can’t buy a fusion welder from Home Depot. The fusion welder equipment is very expensive and only intended for professionals. I thought that maybe I could make my own fusion welder. The professional equipment specs I found on-line called for Teflon plates that could reach 450ºF (230°C), along with some jigs to help align the heater plate between two ends of pipe and then move the heater plate out of the way and press the pipes into alignment.
I started by taking an old toaster apart. I was going to run the elements between two Pyrex glass plates that I found in the cupboard. I figured they were garage sale plates and not part of a set (there were just 3 of them). Boy was I wrong… Those were part of a special 3 plate cake holder thingy that my wife loves. Good thing I checked first.
I decided that her fold-able electric grill would be better because it already has two nice Teflon surfaces and dials for adjusting the temperature. I would just need to break it in half so it would fold outward instead of inward, and probably disable what ever safety switches its designers had included to prevent me from using it that way. Of course, I would also need to buy her a new grill (I already got her a new toaster), but that would be a lot cheaper than hiring a guy with a professional fusion machine, so win-win.
I put rings sliced from my pipe on the grill and the edge softened right away. I then lifted them off and pressed them together… Instant fusion weld… Actually, I guess I heated them too much (too soft) and pressed them together too hard, because I got a bit of a bead inside.
Later, I sliced up my samples, including a cross cut so I could see that the fusion weld was as strong as the rest of the pipe. For scale, the pipe shown in this image is 1/4 inch thick (twice as thick as the pipe I plan to use eventually), so the bead is about 1/12 of an inch. (sorry the pic isn’t very good, my camera doesn’t do macro well, but you can see the bump where the soft plastic at the join pushed into the pipe). Well, that was easy. I am sure I can handle that.
Of course, I wanted to see what else I could use to fuse the plastic… I have a small benzomatic torch. I thought maybe it would burn the plastic, but, even with the direct flame to the plastic, it only burned for a second (some surface residue) and then it just softened the HDPE nicely. The problem was the heat was not even enough… So for an additional experiment, I used the benzomatic to heat a piece of metal and put the plastic against the other side… That distributed the heat well to soften the plastic evenly and wouldn’t require any electricity. For one attempt after the metal was probably too hot, the HDPE plastic did stick to the metal a little, but a piece of my wife’s parchment paper fixed that problem (just like fusing perler beads). I later hooked up my benzomatic hotknife attachment and found I could cut the HDPE pretty well with that.
In general, I found that the HDPE plastic softened easily, once soft, it was a bit tacky to the touch, but would instantly fuse with other HDPE plastic. I found that the joints seemed as solid as the rest of the pipe. I also found that the heavy plastic also kept its heat well (high Specific Heat Capacity), so I had quite a bit of time to get the two pieces together.
The only hard part was aligning the two pipes perfectly. I imagine that would be even more difficult with 20ft long sections of pipe, but I am sure I could build a simple jig to make that alignment much easier.
Quotes (estimates)
I started back up the process of getting quotes last week. I probably called a dozen companies. Only one has got back to me with a quote (so far). A couple others just had follow up questions. And in one case, I am still waiting for a call from the “lady in the office who knows the email and such.” I need the email address to send in the plans.
The one quote that did come in this week was for the footings. It was about 1/4 the price of the last footings quote I got and this guy seemed much more interested in the project and much more pleasant to work with.
Previously, excavators had all told me that they would get down to the depth at the top of the footings and would let the whoever did the footings excavate from there. The other foundation people I spoke to agreed with this and included several thousand dollars of additional excavation in their quotes. However, this latest foundation guy said that it was very difficult to dig a precisely curved trench with their equipment and my sandy site probably wouldn’t be well suited to trench footings anyway. It would be much easier for the excavator to level out the area to the bottom of the trench depth (an extra foot) and then the foundation guy could lay out the curved forms (just thin plywood staked in place) in an open flat space in much less time and much more precisely.
This foundations guy is actually a full service concrete company that also has Shotcrete equipment. It looks like his experience is mostly limited to smaller jobs like turning “michgian basements” into real basements. I still prefer my other shotcrete guy, if I can ever manage to arrange a meeting with him. The foundations guy also said he would do flatwork and gave me reasonable rates for that.
Much Simpler…
I know I have mentioned this tiny house design site before, but I saw another post that I want to share… They have a few small underground homes and even more green roof homes and I recently stumbled on to another one (posted mid 2012) here. Man that look so easy to build compared to mine ;^)
So the year is over. It didn’t exactly go as I had hoped, but we are definitely in a better position to build next year than we were a year ago, more on that in my January kick off post. It is a bit frustrating as I put off doing other things because we expected to build, but such is life. This past month, I took things pretty easy. I only called one person for a quote (a very reasonable sounding plumber that I will bug again in January), and I put some time into building the home virtually.
Virtual construction is a way to run thru the process of building without the huge expense. I individually model each stud, track, buck and ICF piece and then assemble them as I would in the actual construction. Along the way, I am thinking about sequencing and other assembly problems. I try to build to the plan, but make notes when I need to make changes so things will fit. Of course, the computer could make things a bit too easy, for instance, I could just array my studs along and have perfect spacing. Instead, I tried to build things more manually by inserting and positioning each piece.
The below image shows the first steps of my basic construction plan.
I would start by pouring the footing, which includes a lot of precise layout, rebar, form work, etc. Actually, once that footing is done, I will be much more relaxed for the rest of it. I would then set the track (curved or straight) for my steel studs into the top of the footing with Tapcon screws. Then I would place wooden window and door bucks to guide the spacing of the studs. The Marinoware steel studs and steel frame jambs would be set (plumb) and screwed into the tracks. Flat straps are screwed to the outside of the studs to increase the rigidity of the assembly. Metal lath is screwed to the inside of the studs and track, partially to add rigidity, but mostly as a curved backing to catch and form the shotcrete.
- Basic construction detail
Along the way, I spent a lot of time figuring out steel stud placement. I am planning to use “balloon framing” so that load bearing walls will line up and I won’t need to spend as much time or money on laying the curved track. Since the steel studs will just be fancy integral formwork, I don’t need to follow strict structural guidelines. Instead of 2ft or less spacing, I can go a bit wider and not need to cut as many studs for windows, etc. I can vary the placement of the studs to line up with windows and doors across the various levels. For instance, there is an opening directly above the basement opening (shown above). If I shifted the main floor opening by a few inches, I could use one long steel frame jamb instead of needing to place two. I could also align the guest bathroom door with a mezzanine window buck and simplify the framework there also.
After all the steel studs, rebar and conduit are in place, with metal lath to keep things stiff and catch the shotcrete, we would shoot the basement walls 8 inches thick. Since the steel studs are 4 inches deep, this would leave 4 inches beyond the studs. The rebar, just outside of the steel studs, would be nearly in the middle of the wall.
Even though many of the steel studs go thru 3 floors (balloon framing), only the basement shotcrete would be applied in this phase. The top of this shotcrete wall would be cut/troweled roughly level six inches below the floor level. We may even be able to set the column base plates in before the shotcrete sets completely. We would then setup the steel columns and ring beam (I will need a crane of some sort) that will later support my concrete arches. I may do a little MiG welding to tack things in place at this stage.
This assembly looked fine at first, but I later discovered that the ring beam intersected the door buck by several inches. This was a mistake I had not caught earlier. I decided the best fix would be to notch the ribs… These thumbnails illustrate.
The floor system I have chosen uses ICFs (Insulated Concrete Forms) to support the concrete (instead of the wooden subfloor the architect/engineer specified, so yes, I will need to get this stamped by an engineer at some point, I was quoted $800 for that). These QuadDeck ICFs have light steel joists built-in to reduce the number of temporary supports needed. The ends will rest on the shotcrete walls with an inch or two of polystyrene overlap (roughly cut with a hotwire) and be tied in with bent rebar pieces.
This next image is a detail of the rebar (shotcrete and other details are not shown for clarity). Horizontal rebar is wire tied to the outside of the steel studs and should increase the stiffness of the formwork. I considered threading it thru the holes, but that would probably be quite a hassle. The vertical rebar is wired to the inside of the horizontal rebar, but with 8 inches of thickness, there is some flexibility here if needed. Rebar is placed before the shotcrete, and then some pieces are bent down into the ICF channels as needed. More can be added later and bent upward if necessary. I anticipate some issues in certain locations where the QuadDeck spacing clashes with the steel stud spacing, but it should be manageable. I also considered/modeled rebar placement between other ICF sections.
After the ICF forms are placed, they can be walked on. The electrician and plumber should appreciate how easy it is to set conduit and piping into the ICF decking. The ICFs are about 7 inches deep and we will be adding 5 inches of concrete on top of that, so there is plenty of room to work. We also plan to lay radiant flooring across the top of the ICFs. There are little plastic clips that screw into the polystyrene easily to support the radiant tubing. We will screw a form to side of the wall to contain the poor. It will only need to support 12 inches of concrete, so it shouldn’t be a problem.
Some of you may need a cross section vertical to understand the quad deck floor properly… Here it is. Basically, the ICFs form the “negative space” so that the concrete forms a nT (like a single or double-T, but continuous) beam across the spans. Where we meet a wall, the ICF is cut away so the concrete rests on the wall (tied in with rebar not shown here). On the left side of this example, I have cut away some of the ICF in order to create a concrete cross beam above the stairs to the basement. After the concrete is poured, the ICFs remain to provide insulation against heat and noise. They will ensure that my radiant heat goes up.
Now that there is a floor to walk on, additional studs, conduit, bucks, etc. are added in preparation for the main floor shotcrete. Since many studs penetrated from the floor below, this additional work is just incremental. I have not modeled it all yet, but it will include the fill walls under the ribs, the walls around the spiral stairs, etc. Once these walls are up, I will want to model the arches across the donut vault, the hop across the basement stairs, etc.
Here is a final look at how far I got during my Christmas break…
This process has been interesting so far and has resulted in a number of small changes. For instance, I noticed that the original plans placed the mezzanine windows such that the east one was right up against the east wall. When I built the 3D model, I placed the stud so that it would be at the surface of the 8 inch thick wall, but realized that one of the 12 inch thick ribs actually intersected the studs supporting the window. If I wanted to build the balloon wall before placing the ribs, I would need to move the stud by at least 4 inches to clear the rib. I also planned to put in a 14″ sun tunnel thru that corner to bring daylight to the guest bathroom, and that would also have passed in front of the window… So I decided to move all the mezz windows over by 16 inches. Since the steel studs were also framing main floor and basement doors, those would also need to move (to keep the steel stud layout simple). In the computer, this was much much easier to fix than it would have been if I had discovered the issue during the actual construction.
Similarly, I decided to lower the basement egress windows by 6 inches, I added a door way in the basement, etc.
Now I was working on my gantt chart and considering how much time would be taken making these only two at a time. I also considered how much extra cost would be incurred with the multiple crane visits (to set them up 2 at a time). We had already planned to put a slab in-front of the garage (even if the rest of the driveway will be gravel), so the boys could play basketball. I started considering the option of placing the slab earlier in the process and building a few more forms. Even if the extra forms cost an additional 600$ each, I could make 2 more and save on 2 expensive crane visits and a couple weeks of time. I would probably at least break even on money, but save time.
So, I added the ribs to my garage model… You can see that it would be very difficult to fit a 3rd rib on the pad. Yes I tried other configurations, but I need room for the forms around the ribs and space too work, and I couldn’t let them go under the quonset or the crane wouldn’t be able to lift them up without dragging them. However, for the other ribs, I decided to make two left hand ribs and they can be put closer together…
Another option may be that I would use the same glue up styrofoam ceiling tiles that I planned to use to form (impress) the ribs. It may look good to have the same pattern in the vaults between the ribs, but I am a bit concerned about fitting the square pattern to the curved and radiating vault shapes and it wouldn’t work at all in the compound curved bedroom vaults.
