And unwise, I ween. It's harder to get solar heat into water than air.
A low-mass sunspace or solar attic or solar siding and ordinary house mass
can provide all the heat on an average December day as we only store
a small fraction for showers and cloudy day heat in a large cheap tank,
eg a 15' diameter x 4' tall Easy-Set inflatable pool ($44.99 with free
shipping from Amazon :-) in the basement, which might be heated with
a $5 car radiator and its 20 watt fans in the sunspace.
Just thinking about them. Storing heat in the floor means we can't turn it
off on a warm afternoon or save much energy with a night thermostat setback,
and the storage temp can't be more than about 80 F, if we have to "live
inside the heat battery." We can store (120-70)/(80-70) = 5 times more
useful heat in a battery that cools from 120 vs 80 to 70 F, and water
can store about 3X more heat than the same volume of dry sand or 2X more
than wet sand, with a lot less thermal resistance, so it can be efficiently
warmed to a higher temp and supply useful heat at a lower temp, which
increases the effective storage capacity. If we can store 120 F water in
a 15' Easy Set pool, that can provide (120-70)pi(15/2)^2x4x62 = 2.2 million
Btu of useful heat, equivalent to about 22 gallons of oil. Storing the same
heat would require at least 2.2M/(25Btu/ft^3-F(80-70)) = 8.8K ft^3 of slab
concrete, ie 326 cubic yards, eg a 24'x24'x15'-thick slab.
I'm not fond of concrete or burying pipes, but I just got a new concrete
floor in my kitchen (the wood floor over a crawlspace collapsed, after
188 years :-), and the builder put a PEX loop in the concrete, in case
some future owner wants a warm floor. I plan to heat the kitchen with
warm air from the basement. Maybe not much. Stone walls with insulation
outside make the kitchen temp change slowly. One winter I kept it about
36 F by keeping the door to the rest of the house closed most of the time.
If I left a milk jug on the counter overnight, that was OK :-)
Near Soldier's Grove? December is the worst-case month for solar house
heating in Madison, when 460 Btu/ft^2 of sun falls on the ground and
810 (300 diffuse) falls on a south wall on an average 21.7 F day with
a 29.8 high. A really simple 48'x48' house with an R40 ceiling and R30
walls and no windows or air leaks and or electrical energy use that's 70 F
for 12 hours per day and 60 for the other 12 would need 24h(65-21.7)109
= 113K Btu/day, about the same as a gallon of oil...
... 1 ft^2 of R2 polycarbonate glazing with 80% solar transmission might
collect 648 Btu and lose 6h(100F-26F)1ft^2/R2 = 222, for a net gain of
426 Btu/ft^2, so we could warm the house on an average day with 113K/426
= 265 ft^2 of solar siding or 33'x8' of sunspace glazing.
If it's 70 F at dusk and 60 12 hours later, RC = -12/ln(60-21.7)/(70-21.7)
= 52 hours = 109/C, so the house needs C = 5683 Btu/F of thermal mass.
A typical house with furnishings has about 7K Btu/F, 0.5 Btu/F for
1 ft^2 of 1/2" drywall, and so on.
Got numbers? I didn't say sand wouldn't work, but getting close to 100%
solar heat in December seems to require lots and lots of dollars for
insulation (think Michelin man :-) and PEX and sand and hydronic
collectors. If achieving 97% solar heat in December costs 100X more
with sand than water, why use sand?
Maybe it isn't, his way. Then again, PE Norman Saunders estimated that
some of his solar house clients would only have to "purchase heat" once
every 35 years in cold, cloudy, New England.
I figure (with actual numbers :-) the $44.99 Easy-Set pool can store as
much heat as a 24'x24'x15.2' chunk of concrete or a 24'x24'x20' sandbox.
If we only wanted to store 1 million Btu, we might do that in an $0 12'x3'
Easy Set with lots of fiberglass insulation inside the walls and the bottom
and a 10'x10' piece of EPDM rubber attached to a 2x3 frame on top. We need
insulation, and the pool plastic probably can't take 140 F water, which is
a convenient temp for storing lots of solar heat and preheating water for
showers in a $0 1"x300' 13-gallon HDPE black plastic water pipe coil.