Construction - Insulation/Watershed Umbrella for the Annualized GeoSolar System

The insulation/watershed umbrella is a concept advanced by John Hiat in his self-published book, Passive Annual Heat Storage:  Improving the Design of Earth Shelters".

The book is mandatory reading for anyone contemplating a project like ours but it is out of print and apparently available only secondhand online. Don Stephens later improved (in my view) Hiat's PAHS --  including Hiat's umbrella concept, -- and called his iteration "Annualized GeoSolar" (AGS) (also check out  three prior posts on AGS:  first post, second post, third post (the third post zeros in on the insulation/watershed umbrella)).

I began excavating for the umbrella at Thanksgiving-time so it is uncertain how much of it I can do before winter weather interferes.  The umbrella adjacent to the concrete walls will have to wait several months until the walls are insulated on the outside then backfilled to the level of the umbrella. 

Priority Dilemma
My concern is that all of the footings are protected from freezing this winter by completing the umbrella (highly unlikely) or burying the footings under at least 3' of dirt (behind the north wall) or with some other stopgap measure.  As it turns out, I resorted to covering the exposed footings with EPS foam insulation board with the intention of using the boards later to insulate the concrete north wall in the area of the garage.  Unfortunately, the umbrella will have to wait until after the Spring rainy season.  

Actually, a higher priority for what good weather we have left before the ground freezes is to remove the superfluous dirt from in front of the house in order provisionally to restore the pre-construction contour of the slope and bring closer to the underlying original topsoil.  The grading will be finished in conjunction with installing the AGS umbrella in late Spring.

Most of the superfluous dirt came from the final grading for the floor of the
house but considerably more came from excavating for the AGS umbrella.
(Click on the image to enlarge it.)

Superfluous dirt has been moved behind the wall as backfill.

The backfill against the wall is +/- 6' deep in the middle and sloped
towards both ends for surface water drainage; notice  the AGS
conduits protruding upwards; after backfilling against the wall,
the remainder of the dirt from in the front of the house was stacked
to the right of the conduits to a height approximating the final
 backfill level when it is extended all of the way to the wall;
 eventually, the conduits will have to be extended even more in
order to reach daylight through the taller backfill.

Composition of the Insulation/Watershed Umbrella
The umbrella will be merely a "sandwich" made up of plastic sheeting, rigid foam insulation board and sand.  Starting at the bottom in the order in which the layers will be built up, the sequence is as follows:

• Sand over the soil to create a smooth bed on which to lay the umbrella
• 6 mil plastic
• Thin layer of sand
• Mostly expanded polystyrene insulation board; some extruded polystyrene board
• Thin layer of sand 
• 6 mil plastic
• Thick layer of sand
• 6 mil plastic
• Thin layer of sand
• Two layers of recycled synthetic carpet upside down
• Topsoil

I will be layering the plastic and insulation precisely as Hiat recommends except, of course, we have no earth sheltering on the roof.  Stephens' influence will be represented by the carpet overlay to protect the sandwich from

Illustration from Hiat's book (click on picture to magnify the details)

mechanical damage from such things as burrowing critters or thoughtless use of shovels and other tools.  It will also protect somewhat against the penetration of plant roots although the latter will mostly be inhibited by the fact that the only area that is not too dry to support them will be above the first layer of plastic.

As recommended by Hiat, the insulation will be thickest near the house and thinnest at the periphery in five steps corresponding to the width of 4' x 8' foam board --   4" thick for 8 ft,, 3" for 4 ft,  2" for 4 ft and 1" for 4 ft -- making the umbrella 20' wide. Where a single sheet of plastic is not at least 20' wide abutting sheets will be shingled with a large overlap. The sand is important for a couple of reasons.  First, to provide smooth interfaces between the layers and eliminate air spaces that might allow the umbrella to be damaged by heavy equipment or vehicular traffic, particularly punctures of multiple layers of plastic at the same place allowing through-and-through leakage. Second, to hold the plastic sheeting apart enough that any water finding its way through one layer is sure to flow downhill through the porous sand to the periphery.  Without the sand, the backfill over the sandwich would compress the plastic layers together and trap water, particularly where the plastic extends beyond the insulation.  The umbrella will slope away from the house in all directions and, in some places, end in a mini-French drain for faster drainage ("drainage gravel" in the illustration).

The white expanded polystyrene (EPS) foam board will suffice for all areas except under the driveway and the garage floor where pink 250 psi extruded polystyrene (XPS) will be a better choice for supporting vehicles.  The sand in the sandwich will go a long way towards stabilizing and supporting the weaker EPS. The two reasons for selecting EPS for most of the umbrella insulation are price and EPS's long history of use in wet environments (think insulated concrete forms and flotation for boat docks). I trust that the plastic sheets will keep the foam pretty dry but there will be a few small leaks inadvertently created during installation that will admit trickles of water through a given layer that will have to travel in association with the foam and sand layers to an exit at the periphery of the next layer of plastic, so it is prudent to anticipate some exposure to moisture.

The reason for the plastic is to keep the soil under the umbrella dry.  Doing so lowers its thermal conductivity and inhibits transfer of heat from the thermal mass to the outside environment.  Through dry soil, it takes 6 mos for a unit of heat to move 20' which is the basis for making the umbrella 20' wide.  As Hiat explains, even if some moisture would somehow find its way through three layers of plastic, the negative impact on the ground below would be inconsequential, that the saturation would be so spotty and limited that it would not be enough to compromise the efficacy of the AGS system.

Garage Serves As Part of the Umbrella

Insulation of the wide footing is completed by the addition
of  XPS foam board horizontally; it was covered with sand
before the backfilling; EPS would have worked as well

The entire floor of the garage will be insulated for a couple of reasons. One is to complement the other measures that will make the garage warmer, viz., the insulated foundation, the thick walls and ceiling filled with insulation and the insulated overhead doors. The other is to serve as the umbrella for the north half of the east end of the house.  

In order to fulfill the requirements for a frost-protected shallow foundation, the garage foundation footing below the insulated concrete forms will have to be insulated where it is not protected by the house and garage floors or the umbrella next to the south half of the east end of the house. Likewise, the wide footing for the east end of the concrete wall was poured against vertical insulation but it needed to be insulated on top, which was simply a matter of fitting foam board over it, sandwiching it between two thin layers of sand (to carry any water from behind the damp proofing membrane to the footing drain) and backfilling over it. Eventually, a retaining wall will be resting on it.  

Horizontal insulation in place and the backfilling started

I insulated the narrow footing under the garage foundation in conjunction with extending the footing drain to daylight downhill. I enlarged the overdig next to the footing so that it would accomodate foam board 3' wide and embedded the drain in sand, much like along the backside of the wall, screeded the sand level with the top of the footing and laid in the insulation. I did the same procedure for that part of the front footing for the garage that will not be protected by the umbrella, except there was no drain to worry about and the EPS was only 2' wide since it faced south and direct sunshine should warm the backfill enough to make for a warmer footing anyhow.

The footing drain at the other end of the house merely needed to be extended downhill to daylight, bedded in sand and backfilled.  

The extension of the footing drain for the west concrete wall.

With respect to the exposed footing inside the garage that will be covered by the floor eventually, I used soil to fill the overdig flush with the floor grade and laid foam boards against the north wall and the insulated concrete forms for the other three walls and weighted them down with stones. That should protect the footing this winter; the floor will go in next summer.

For some reason, the follow-up post on construction of the insulation/watershead umbrella was posted out of sequence and is dated 11/28/15.  Please drop down four or five posts to find it.

Design - Exterior Walls

Priorities

Being on a strict budget will precipitate a lot of compromises.  However, one issue is not negotiable and that is a super-tight and super-insulated envelope for the house.   Probably this means we will have laminate or wood counter tops instead of natural or man-made stone, or site-made interior doors instead of prehung, but that's okay.  Energy conservation will remain our highest priority.

Stick-Built Walls Using Salvaged Lumber

In a perfect world, we would be using structural insulated panels or insulated concrete forms for the exterior walls but they far exceeded our budget and are limited on  the amount of insulation they can provide.  And we have a lot of free
salvaged dimension lumber that would go to waste if it were not used for the exterior walls.  The trouble with salvaged 2 x 4s however is that the nails often split the ends during the salvage process, especially when toe-nailed.  Typically, a couple of inches must be sacrificed from one or both ends of the boards during the denailing process, resulting in many studs that are too short for typical 8' wall construction.

Wall Trusses

Early on, I decided to use wall trusses instead of individual two-bys as "studs" for several important reasons.  One, with trusses, the salvaged 2 x 4s could be utilized for an 8' walleven if shortened as long as they remained at least 93" long.  Two, some of the 2 x 4s were not entirely straight (but neither is a lot of new lumber these days) but could still be utilized quite well for trusses.  Third, trusses would essentially eliminate thermal bridging. The last and most important reason for using trusses is that R-value for the walls is directly proportional to the thickness of the insulation and trusses can be designed to house whatever thickness of insulation needed to hit a given R-value target.

Truss Design

The 2 x 4s in the trusses will be turned 90 degrees from the way a stud normally sits in a wall and arranged in pairs, one facing outward and the other inward.  They will be tied together with short  2 x 4s at the top and the bottom and braced with three pairs of gussets cut from 3/8" or 1/2" plywood or OSB.  When arranged in the wall on 24" centers, the result will be a wall filled with 15" of insulation and virtually no thermal bridging.  What bridging does occur will be limited to the short 2 x 4s at the tops and bottoms of the trusses and through the skinny gussets.

As far as straightening the boards, my original idea was to assemble them at the tops and bottoms with the bows pointed away from each other.  Then I planned to use clamps to pull them together until they fit the precut gussets.  Once the gussets were attached, the truss would be straight.  

Using Jigs
However, while researching rice hulls as insulation, I came upon a most interesting

Wall truss jig

slide presentation  (Rice hull house) showing the use of wall trusses similar to what I envisioned but being built in jigs for ease of construction and for standardization.  At the time of this writing, a jig was already in service (the subject of another post) that facilitated any straightening that needed to be done.

Top and Bottom Plates
Characteristically the top and bottom plates are as wide as the wall is thick, e.g., 2 x 4s for 3 1/2" walls and 2 x 6s for 5 1/2" walls.  Unfortunately, the plates then become conductive thermal bridges because they are exposed simultaneously to both the exterior and interior environments.  In order to solve this problem, our 15" walls will have double 2 x 6 mudsills side-by-side and  2 x 6 top plates that are double in a side-by-side sense as well as two courses on top of each other in the typical fashion.  The side-by-side configuration of the mudsills and the top plates will allow 4" of insulation between them, thereby arresting thermal bridging. 

Rice Hull Insulation
As discussed in a prior post, we plan to use rice hull insulation, a cheaper and greener

alternative to conventional insulation. With 15" walls, rice hull insulation at R-3.2 per inch, we expect easily to exceed an R-factor of 45 and, with such minimal thermal bridging with the trusses, probably more like the higher '40s.

The logistics for transporting the hulls from the Mississippi delta area and getting them into the walls is still in flux. My tentative plan is to keep the garage un-roofed until the hulls have been dumped into it then use a homemade blower patterned after the one shown in the above link to the Rice hull house to get the hulls into the walls (and ceilings).  More on this later.

Design - Windows and Doors

Even with our tight budget, we intend not to compromise on window and door quality because they are the weak link in heat transfer in and out of the building envelope.  For example, the R-value of double pane glass is a paltry 1.5 - 2 while the R-value for our exterior walls will be over 45.  The best way we can maximize the thermal performance of the windows and doors is by controlling air infiltration between the wall framing and the window.and between the window frame and the window sash.  Then we need to control thermal bridging through the window frames and through the glass itself.

Air Infiltration
The two ways air infiltrates windows is between the window and the framing and between the window sash and the frame.  In order to hold the window back from the plane of the exterior wall (see below), we will have to use replacement windows that do not have nailing flanges like that of new construction windows.  Nailing flanges are the first line of defense against air leakage between the window frame and the framing.  We will have to compensate for the lack of flanges with caulking and minimal expanding foam insulation and precise fitting of the inside and outside trim.


Air infiltration between the frame and the sash is easier to fix.  Instead of sliding windows

Casement window

(single- or double-hung or horizontal sliders) we will use hinged windows that close against and compress a semi-rigid air seal that doesn't leak. Sliding windows are built with more tolerance between the window and the frame in order to facilitate sliding.  The air sealing is done with flexible air seals that tend not to be as airtight. For us, all but a couple of windows in the garage will be casement windows.

Thermal Bridging

Thermal bridging, a form of conductive heat transfer, occurs through both the glass and the frames.  Double glass marginally improves the R-factor of single pane glass by providing a dead air space between the panes.  An R-2 is not much to write home about. Since windows are essentially poorly plugged holes in the wall, a better strategy is to limit the number and size of windows, particularly on the north and west sides of the house (US Midwest).  Except for two or three small windows on the east, two of which are in the garage, all of our windows will face south.

The addition of argon gas between panes of glass also helps to reduce thermal bridging by impeding convection currents in the space between panes. The literature in my library and online information lead me to believe that argon is more hype than help because it escapes within a few years and therefore is not a good investment.  Recently, the sales rep for the window company that will supply our windows convinced me otherwise. The argon in his windows is guaranteed to be 80% effective for the first 20 years and costs only about $10 per window.  That sounds like a good investment after all.

Thermal bridging through the frame is easier to control than bridging through the glass by choosing materials that have low thermal conductivity.  Fiberglass and vinyl are better insulators than wood and wood is better than metal so, if our budget allows, we will opt for fiberglass for both windows and doors.  If not, vinyl will be a reluctant second choice at least for windows.  Nevertheless, choosing vinyl would be a double-edged sword from a green building standpoint; while it  minimizes thermal bridging it comes largely from petroleum. Fiberglass is a green choice for three additional reasons:  (a) it has the same coefficient of thermal expansion

Proprietary graphic comparing coefficient of thermal expansion 

as glass so the seal between the panes and between the panes and the frame lasts longer, (b) it is many times stronger than

Another proprietary graphic showing the strength of fiberglass relative to vinyl 

vinyl and (c) it is greener because it is made from the most abundant resource on earth -- sand. Also, we will specify warm edge spacers between the panes, which, at nominal cost, reduces thermal bridging through the edges of the glass.  However, we will not opt for fully insulated frames because the payback is not as quick here as in northern climates.

Winter Winds
Glass that directly faces winter winds is considerably more conductive than glass on the leeward side of the house because the wind removes the thin layer of insulating air on the exterior surface of the glass (wind-washing), thereby accelerating the loss of "fresh" heat through the glass.  We plan no windows directly facing the prevailing winter winds. Our exterior walls will be 15+ inches thick which will allow us to recess the south-facing windows into the wall as much as 6" thus sheltering them from wind-washing more than if they were mounted flush with the wall.

The fact that the house will be earth sheltered on the north side and half of the west side will pretty much neutralize cold winter winds.  Nevertheless, soon after purchasing our property years ago, we began establishing a shelter belt north and west of the future house as another buffer.  And the eastern red cedars that we planted are native to our area and very beneficial for wildlife. 

Tinting and Low-E

Exterior overhangs will shade the windows from the summer sun but will allow winter solar gain.  The gain will be maximized by using clear glass instead of tinted glass. However, we will specify a low-E coating in order to slow heat loss out through the glass in winter. Admittedly, low-E will diminish slightly the amount of solar gain in winter but our AGS system is the principal heat source and gain through the windows will be welcome but not essential.

That said, we may leave off the second level overhangs for the first one or two years.  It will probably take that long for the AGS system to charge the thermal mass with enough heat for a year-round constant floating temperature at a comfortable level.  Accepting some summer solar gain to augment the output from the solar collector might be a good strategy.

Also, after the AGS system reaches equilibrium, we will assess the need for thermal window shades for nighttime and gray day use.  If the system provides plenty of heat, they would be superfluous, if not, they can be added.  Their function would be to keep moving air (convection) away from the glass and thereby slow conductive heat loss through the glass.  Secondarily, less convection would make for a more comfortable living environment.  In order to be successful, the shades would need to be sealed as tightly as possible on all four edges.  Heavy drapes, regular window shades or louvered blinds help some but are not the same thing and, in some cases, actually enhance convection.