re: sod on a shed/pole roof--will it work???
19 dec 2002
>this upcoming year i'm planning to build a small (16' wide x 12' deep)
>passive solar 'cabin' for weekend and summer use. the general design
>will be an 'inverse lean-to', with the rear northwest-facing wall
>being 9' high tapering to a 7.5' front southeast-facing wall.
smallest wall to the south? you might collect solar heat with a steeper roof.
>this 'inverse lean-to' design will (i hope) have two advantages. first, it
>will channel water down and off the front of the structure (since it
>is build on a slight hill facing downhill), rather than towards the
>rear, where it has to be dealt with or risk having it find it's way
>into the structure. second, this design will mean that the high point
>of the ceiling will be at the rear of the structure, where the
>stovepipe chimney will be located...
some people define solar cabins as "those with no other form of heat."
>the side and rear walls of this structure will be 'earthbag
>construction' (14" x 24" polypro bags filled with onsite soil, laid in
>alternating courses like the old sod houses of the great plains,
jebby, hitch up the horse and gimme one a them 14" x 24" polypro bags...
>with barbed-wire between courses, and finished with cement stucco).
will it stick?
>since every 3rd or 4th course the bags are laid 'lengthwise' to stabilize
>the underlying rows, these final walls will end up being around 24" thick
>(excluding blueboard insulation and stucco) and very strong.
>my main question/concern (aside from any other suggestions anyone
>might have on my general design) concerns the roof. i'd like to keep
>it simple with a traditional lumber 'open-faced shed style' roof.
>this style of roof, properly insulated...
>and finished with some precut corrugated steel roofing panels, would be
>easy and dependable. if i stuck with this sort of roof, i would probably
>also go with a straightforward lumber south-facing front wall. lumber
>rather than earthbags would probably be easier to work with as far as
>mounting the two large (4' x 5') windows for the direct solar gain on
>this front wall...
ah. dimensions... 40 ft^2 of r2 south windows with 80% solar transmission?
and the weather? would 1000 btu/ft^2 fall on a south wall on an average 30 f
january day, as in phila? you might check out "ohm's law for heatflow."
>...thermal mass in my floor and walls will be sufficient enough to offset
>the loss of using lumber for the front wall (assuming proper insulation).
thermal mass is not a substitute for insulation, but it can be good, if
it has insulation on the outside and lots of surface exposed to room air
on the inside. page 97 of d.c. beard's 1914 book "shelters, shacks &
shanties" has a sketch of a shack with walls made from barrels stacked
2-high, with flat boards on top of the barrels. he writes:
no observing person has travelled far upon the american railroads
without noticing, alongside the tracks, the queer little [sod?]
houses built of railroad ties... figure 142 shows a barrel dugout.
it is made by digging a place for it in the bank and, after the floor
is levelled off, setting rows of barrels around the foundation, filling
these barrels with sand, gravel, or dirt, then placing another row
on top of the first, leaving spaces for a window and a door...
after which the walls are roofed with logs and covered with sod...
the dirt is next filled around the sides, except at the window
opening... a barrel also does duty as a chimney.
shacks like this are used by homesteaders, miners, trappers, and
hunters; in fact, these people use any sort of material they have
at hand. when a mining-camp is near by the freight wagons are
constantly bringing in supplies, and these supplies are done up
in packages of some kind. boards are frequently worth more a yard
than silk, or were in the olden days, and so the home builders
used other material. they built themselves houses of discarded beer
bottles, of kerosene cans, of packing boxes, of any and every thing.
usually these houses were dugouts, as is the barrel one...
in the big tree country they not infrequently made a house of
a hollow stump of a large redwood, and one stone-mason hollowed out
a huge bowlder for his dwelling, but such shacks belong among the
freak shelters. the barrel one, however, being the more practical
and one that can be used almost anywhere where timber is scarce
but where goods are transported in barrels, deserves a place here
amoung our shacks, shelters, and shanties...
a 55 gallon drum is about 3' tall and 2' in diameter, with about 25 ft^2
of surface and 450 btu/f of thermal mass if filled with water. they can
be free. i collected about 300 free from a local turkey packing plant.
waterwalls would be an alternative... shelves about 1' wide made from 2x4s
and bolts and 2"x4" welded wire fencing, holding water-filled plastic film
55 gallon drum liners and blueboard on the outside, with latex paint.
>however, ideally i'd like to have a sod roof, both for aesthetic and
>functional reasons (slightly more insulation value and evaporative
>cooling benefit during the summer). but i'm a little concerned about
>whether the roof i've described would be sufficiently strong to hold a
you have left out a few details. you may not need evaporative cooling if
your cabin has enough thermal mass and insulation and night ventilation.
>...i've seen figures ranging from a minimum of 4" necessary to a maximum of
>18". i came across some figures ballparking the weight of soil at bewteen
>90-110 lbs per cubic foot. that's obviously 100+ pounds per square
>foot for a 12" layer of sod, and of course you have to keep in mind
>additional weight when wet, snow load in winter (esp. with the low
>roof slope i'm talking about here) and the weight of the other roofing
>materials. we don't get a lot of wet snow where i'm at (avg 16"
>precip year), but every once in awhile we get hit with a pretty good
>blizzard. the way the house is configured much of the snow would blow
>over the house, drifting in the front, but still there is to be
>expected some snow load from time to time.
the boca code would have the snow load requirement for your location.
>i'm wondering whether if i modified the south-facing front wall to be
>an earthbag wall that would provide sufficient support for a sod roof?
you might measure the tensile strength of a strip of the bag material.
> i would also buttress the side and front walls a little (as earthbag
>builders suggest when you build a vertical, rather than domed wall).
domed sounds better. maybe a simple timber dome covered with chicken wire,
then poly film, then a tire hairnet filled with soil for mass, then more
poly film, another tire hairnet filled with leaves for insulation, more
poly film, and another tire hairnet filled with peat moss to provide a home
for suitable volunteer roof vegetation.
>'any ideas/thoughts/suggestions greatly appreciated!
how about a barrel shack with a shallow 12' wide roofpond made from a
single 20' wide piece of epdm rubber? you might add a solar pool cover
in cold weather. make the whole south wall a shallow poly film sunspace
over blueboard, with an air gap. in phila, a square foot of r1 sunspace
glazing with 90% solar transmission would collect 0.9x1000 = 900 btu and
lose about 6h(80-30)1ft^2/r1 = 300 on an average january day, for a net
gain of 8x16x600 = 76.8k btu. with 52 drums stacked 2-high and 12'x16' id
and 16x20' od and 896 ft^2 of exterior surface and lots of thermal mass
inside, 24h(65-30)896/r = 76.8k in january in phila makes r = 9.8, if we
ignore the warmer roof and the warmer south wall during the day.
with 2" (r10) blueboard walls and ceiling, rc = 10/896x104x450 = 522 hours.
if the cabin were 65 f at the end of a long string of average days, it would
cool to 30+(65-30)e^(-5x24/522) = 57.8 f after 5 30 f cloudy days in a row.
the 25x104 = 2600 ft^2 of drum surface would have about 3900 btu/h-f of
conductance to room air. the drums need to store 18h/24hx76.8k/6h = 9600
btu/h from warm sunspace air on an average day, which would raise the room
air temp by 3.3 f to 68.3 f. the drum water temp would increase by 1.2 f.
at night, we would have something like this:
1/3900 | 10/896
65 ---www---*---www--- 30 i = (65-30)/0.0114 = 3066 btu/h,
i so t = 65-3066/3900 = 64.1 f.