Consumer Homes.

Question:

Hi, I’d like to ask a few questions, first, about the solar possibilites of a geodesic dome home.  Has the solar closet been modified for such a structure? I am assuming the structure will consist of a dome about 28 ft in diameter, supported on a riser wall of 3 ft. Also, After visiting this site http://www.ibiblio.org/ecolandtech/links/start-392001/msg00413.html ( The Thermal Cistern and the Solar Closet ) I am wondering about the effectiveness of a solar closet in Canada (say, 43-45n lat ), and the comparative feasibility of a thermal cistern. Thirdly, is this tracking solar concentrator http://www.ida.net/users/tetonsl/solar/solarhom.htm presently the best choice among concentrators for a DIY type, and can anyone point me to other examples with mods, etc? It looks like it would match well with Harry Thomason’s 2 tank design. Fourthly, I’d like to be able to figure out the "thermal mathematics" of a structure, just like Nick Pine does. How impudent! Alec

Response:

>…a geodesic dome home.  Has the solar closet been modified for such a >structure? I am assuming the structure will consist of a dome about 28 ft >in diameter, supported on a riser wall of 3 ft.

A 28′ hemisphere with 2Pi14^2 = 1232 ft^2 of surface on top of a 3′ cylinder with 28Pi3 = 264 ft^2, totaling 1496 ft^2? Or floating on a 28′x4′ deep swimming pool, tracking the sun with a system of ropes and pulleys attached to sunflowers in a sunspace >…After visiting this site >http://www.ibiblio.org/ecolandtech/links/start-392001/msg00413.html ( The >Thermal Cistern and the Solar Closet ) I am wondering about the >effectiveness of a solar closet in Canada (say, 43-45n lat ), and the >comparative feasibility of a thermal cistern.

"Thermal cisterns" seem like poor performers, with too much thermal resistance in the sand, which stores less heat than water by volume and is spread out and hard to insulate, compared to a box with a lower surface-to-volume ratio, and it’s harder to get solar heat into water than air. They would work better with water flowing through the sand itself rather than water in pipes. This needs a waterproof "tank," but the sand could support a wood floor over a vapor barrier, if the water isn’t too deep. How would we heat and insulate it and distribute the heat as needed? John Paul Jones of SECO (the great^N grandson of THE John Paul Jones) suggested something like this 30 years ago, with 70 F water circulating among large sand planters and heat pumps made from water- cooler compressors in every room. The Solarium Workbook (NRC, 1981) has long-term average weather data for                            Dec               Jan                         temp SDD  sun     temp SDD  sun Fredericton, NB      -6 C  82   2127   -9 C  97   2811 Wh/m^2-day Moncton, NB          -5         1895   -8    67   1888 Halifax, NS          -1    79   1650   -3   102   2341 Charlottetown, PEI   -4    77   1838   -7   108   2914 >I’d like to be able to figure out the "thermal mathematics" of a structure, >just like Nick Pine does.

He often starts by finding the "worst-case month" for solar house heating, with the least "sun per degree-day," a metric from Norman Saunders, PE, who divides the amount of solar energy that falls on a south wall on an average day of the month by the indoor-outdoor temp diff, eg SDD = 1888/(20-(-8)) = 67 for Moncton in January, the worst-case combination above. December in Moncton is clearly easier, with more sun and a higher outdoor temperature. Then we convert to quaint US units, from -8 C to 18 F and 1888 Wh/m^2-day to 598 Btu/ft^2-day, and realize this is very cold and cloudy, not a nice climate for solar house heating. Then we imagine the dome has two floors and an attic and 96 ft^2 of R4 windows with 50% solar transmission, with 48 ft^2 facing south and 24 east and 12 north and west. If the dome itself has R30ish insulation, eg 2" Styrofoam outside 6" fiberglass, or 8" SIPs, or plywood panels made from 9.5" I-joists with cellulose fill, so its thermal conductance is 96ft^2/R4 = 24 Btu/h-F (24 buhfs) for the windows plus 1400/30 = 47 for the "walls" plus 30 buhfs for 30 cfm of air leaks (with 30cfmx60/7594ft^3 = 0.24 air changes per hour, a very tight structure), totaling about 100, so it needs about 24h(65-18)100 = 113K Btu on an average January day. A frugal 600 kWh/mo of indoor electrical use could provide 20 kWh/day (68.2K Btu/day) of this. It needs 5(113K-68.2K) = 224K Btu for 5 18 F cloudy days. The Solarium Workbook says a horizontal surface receives 1164 Wh/m^2-day (369 Btu/ft^2-day) of sun, and east and west windows get 908 Wh (288 Btu) in January in Moncton, so the windows collect about 0.5×12(4×598+3×288) = 19.5K Btu on an average day. Subtracting this and the electric heat from the daily heat requirement, we need an additional 113K-68.2K-19.5K = 25K Btu/day of solar heat. This might come from thermosyphoning air heating panels, eg dome triangles with polycarbonate glazing over an air gap and dark window screen as a mesh collector with 70 F air near the glazing and R30 insulation behind the screen, or a transparent skirt over the southern part of the 3′ cylinder. A square foot of R2 air heater glazing with 80% solar transmission might gain 478 Btu over 6 hours on an average day. With 70 F air near the glazing, it might lose about 6h(70-18)1ft^2/R2 = 156 Btu, for a net gain of 332 Btu (not much), so the extra heat might come from 25K/332 = 78 ft^2 of air heaters, eg a 4′x25′ skirt around a 4′ stemwall, if the floor is 4′ above the ground. Higher heaters might avoid reverse thermosyphoning at night with automatic foundation vents, or thermostats and motorized dampers. On an average day, the dome needs to store about 18h/24h(113K-19.5K) = 70K Btu of overnight heat. With a 70-60 = 10 F day/night temp swing, we might store this heat in a 7K Btu/F thermal capacitor with a low series resistance, eg 1400 hollow 5 Btu/F concrete blocks in an 8′x8′x16′ column or a 4′x4′x16′ column with 4" thinwall PVC water pipes threaded through holes in 561 blocks. As an alternative, we might let 100 F air-heater air pool under the first floor ceiling, with something like a 100-80 = 20 F temp swing and 3500 Btu/F (55 ft^3) of water in poly film ducts in the ceiling. If the ceiling has Pi(14^2-5^2) = 527 ft^2 of surface, the water would be 1.2" deep. We might store cloudy-day heat in an 224K/(130-80) = 4480 Btu/F (70 ft^3) well-insulated water tank that cools from 130-80 F over 5 cloudy days and provides hot water, with the help of a 97% greywater heat exchanger. With a 4′ stemwall, the attic floor could have Pi(14^2-12^2) = 163 ft^2 of surface. A draindown floor under an R2 roof with 80% solar transmission might gather 0.8×163x369 = 48K Btu/day of sun and lose 6h(T-18)163ft^2/R2, which makes T = 116 F, or more, since daytime average temps are warmer than 24-hour temps, and in cloudy climates, most of the Btus arrive in a few hours per week of beam sun, "gift-wrapped" for minimum heat loss and possible concentration, according to PE Howard Reichmuth. You can find more thermal math in the 2000-page archive at http://www.ece.vill.edu/~nick, which also has a PDF of the Nov/Dec 2003 Solar Today story "Soldier’s Grove Soldiers On." Nick    It’s a snap to save energy in this country. As soon as more people    become involved in the basic math of heat transfer and get a gut-level,    as well as intellectual, grasp on how a house works, solution after    solution will appear.                                      Tom Smith, 1980

Response:

>A square foot of R2 air heater glazing with 80% solar transmission might gain >478 Btu over 6 hours on an average day. With 70 F air near the glazing, it >might lose about 6h(70-18)1ft^2/R2 = 156 Btu, for a net gain of 332 Btu (not >much), so the extra heat might come from 25K/332 = 78 ft^2 of air heaters, eg >a 4′x25′ skirt around a 4′ stemwall, if the floor is 4′ above the ground. >Higher heaters might avoid reverse thermosyphoning at night with automatic >foundation vents, or thermostats and motorized dampers.

Or David Delaney’s flow separation device, with venting of any excess heat to the outdoors. Nick

Response:

>[snip] >The Solarium Workbook says a horizontal surface receives 1164 Wh/m^2-day >(369 Btu/ft^2-day) of sun, and east and west windows get 908 Wh (288 Btu) in >January in Moncton, so the windows collect about 0.5×12(4×598+3×288) = 19.5K >Btu on an average day. Subtracting this and the electric heat from the daily >heat requirement, we need an additional 113K-68.2K-19.5K = 25K Btu/day of >solar heat. This might come from thermosyphoning air heating panels, eg dome >triangles with polycarbonate glazing over an air gap and dark window screen >as a mesh collector with 70 F air near the glazing and R30 insulation behind >the screen, or a transparent skirt over the southern part of the 3′ cylinder. >[snip]

            I got lost on the seemingly low numbers of 19.5 K BTU/day and 25K Btu/day in Canada, maybe I don’t realize the 100K BTU/HOUR furnaces don’t run very many  minutes per day.             And if it is to be passive solar heated, why would’nt there be more south facing windows and no north facing windows.             I don’t see where a geodesic dome would be good for passive solar heat design, and the living areas should be designed to take advantage of the times the sun shines through certain windows or air heater panels.             Breakfast should be in the southeast room, midday work or lounging should be in the central south facing room, and evening activities in a southwest facing room, all these rooms should have doors so they can be isolated when cold and not be used.            The north wall should be all storage closets or utility rooms that don’t need to be heated, and they would moderate the heat loss to the north wall.              One problem is having direct sun in the room, it is bad for eyes, and will damage wood and fabric furniture.            My next door neighbor bought a manufactured house built in New England, all electric, and the biggest electric bill has been $150 for the worse months.            I don’t know if the have a heat pump or not, but using gas, my small old house uses $250 in gas and electric in the worse months, and gas is going up at least 10 percent.                It seems like a hundred cubic feet of gas is a nominal 100K BTU, and costs me about $1 now, while 30KWH costs me $1.80 for about the same amount of electric energy.            That is why 25K BTU/day (= 8Kwh/day) seems like too small an amount to worry about.            I can shut off the entire north side of the house, there is no water there, and reduce energy use if I want.    But if a house is designed for passive solar, north facing windows will not be missed, and south and west facing windows may need overhangs or shields to avoid excessive heat in summer. Joe Fischer

Response:

>>The Solarium Workbook says a horizontal surface receives 1164 Wh/m^2-day >(369 Btu/ft^2-day) of sun, and east and west windows get 908 Wh (288 Btu) > in January in Moncton, so the windows collect about 0.5×12(4×598+3×288) >= 19.5K Btu on an average day… >I got lost on the seemingly low numbers of 19.5 K BTU/day and 25K Btu/day >in Canada…

Not much sun, in that cold climate. >And if it is to be passive solar heated, why would’nt there >be more south facing windows and no north facing windows.

There are a reasonable number of each, to reduce glare, and some air heaters that lose no heat on cloudy days or at night. Nick

Response:

>…a geodesic dome home.  Has the solar closet been modified for such a >structure? I am assuming the structure will consist of a dome about 28 ft >in diameter, supported on a riser wall of 3 ft. > A 28′ hemisphere with 2Pi14^2 = 1232 ft^2 of surface on top of a 3′ cylinder > with 28Pi3 = 264 ft^2, totaling 1496 ft^2? Or floating on a 28′x4′ deep > swimming pool, tracking the sun with a system of ropes and pulleys attached > to sunflowers in a sunspace

No no, I envision a "floating" house, but a la Larry Walters: http://www.markbarry.com/ ( click on "The Lawn Chair Pilot" on the right side ) But seriously, I am thinking of using concrete piles at the vertices for footings, with a wooden main floor, no basement. Being someone who just can’t seem to comprehend without the crutch of a picture or two, I’ve modelled what I think is what you mean… http://www.chebucto.ns.ca/~ba325/solar.html – Hide quoted text — Show quoted text -> "Thermal cisterns" seem like poor performers, with too much thermal resistance > in the sand, which stores less heat than water by volume and is spread out and > hard to insulate, compared to a box with a lower surface-to-volume ratio, and > it’s harder to get solar heat into water than air. They would work better with > water flowing through the sand itself rather than water in pipes. This needs > a waterproof "tank," but the sand could support a wood floor over a vapor > barrier, if the water isn’t too deep. How would we heat and insulate it and > distribute the heat as needed? John Paul Jones of SECO (the great^N grandson > of THE John Paul Jones) suggested something like this 30 years ago, with 70 F > water circulating among large sand planters and heat pumps made from water- > cooler compressors in every room.

Fortunately, "I have not yet begun to build!". Alec

Response:

>I am thinking of using concrete piles at the vertices for footings, with >a wooden main floor… I’ve modelled what I think is what you mean…

http://www.chebucto.ns.ca/~ba325/solar.html Nice drawing, with the air heaters in the right place, but it looks like the dome starts 7′ vs 3′ or 4′ above the ground and the concrete block "column" is turned sideways as a shoebox under the floor, and there’s no solar attic. A column might operate with natural thermosyphoning air flow, with no fans or blowers, absorbing heat in the afternoon as warm air near the top flows into the column and cools and sinks, and providing heat in the morning as cool room air near the bottom enters the column near the bottom and warms as it rises and exits near the top. But water is cheaper than concrete and stores more heat by volume, often with less thermal resistance, and a warmer ceiling with a larger temp swing requires less mass to store the same heat, so why not put about 1" of water in poly film ducts under the first floor, on top of a plywood ceiling? You might drape another layer of poly film over the ceiling first, to avoid leaks. With a central post and 10 14′ radial beams and a 4′ stemwall and a Pi(14^2-4^2) = 565 ft^2 ceiling, each wedge only needs to support 3500/10 = 350 pounds of water, with M = WL/8 = 350×14x12/8 = 7350 in-lb and S = M/f = 7.35 and d = sqrt(6S/b) = sqrt(44.1/1.5) = 5.4", so it looks like you can do this with 2×6s, ignoring the plywood strength. You might store cloudy-day heat in a 70 ft^3 ~4′ cubical tank under the floor that cools from 130-80 F over 5 cloudy days, and warm the dome on a cloudy day by pumping some water up through the ceiling. The attic floor needn’t contain much water, just enough for solar collection during the day. It would have insulation underneath, and maybe a few transparent patches for daylighting. With separate average and cloudy day tanks, you might avoid the air heaters. Nick

Response:

> …but it looks like > the dome starts 7′ vs 3′ or 4′ above the ground and the concrete block > "column" is turned sideways as a shoebox under the floor, and there’s no > solar attic. > A column might operate with natural thermosyphoning air flow…

So the concrete column ( which I failed to realize…got stuck on the 4′ dimension ) acts as a thermal "chimney". I’m starting to like the water ideas anyhow. > But water is cheaper than concrete and stores more heat by volume… > You might store cloudy-day heat in a 70 ft^3 ~4′ cubical tank under the floor > that cools from 130-80 F over 5 cloudy days, and warm the dome on a cloudy day > by pumping some water up through the ceiling. The attic floor needn’t contain > much water, just enough for solar collection during the day. It would have > insulation underneath, and maybe a few transparent patches for daylighting. > With separate average and cloudy day tanks, you might avoid the air

heaters. First, I’d like to peg down the possible methods/devices for this setup, so I can focus on a more specific design. It’s certainly a case of resizing and reshaping the egg to accomodate an ever-changing chick! I should have made things more clear from the start as to my particular desires. For one, I’d like to have a floor plan which only includes a semicircle for the second floor, thus opening up the visual and air flow space. Something like: http://www.chebucto.ns.ca/~ba325/floorplan.html. If that is workable without drastically changing things, are these some of my possibilities: 1) One or Two separate, one-for-cloudy-day ( stored under the first floor? ), water heat storage units in the form of 2 Thomason’s-pressurized-tank-within-nonpressurized-box units being heated by:     a) Solar air heaters in a 4′ stem wall, below the main floor, heating an under-second-floor water layer by trapping the hot air which is then circulated by pump to the tank(s);     Or     b) The same but with the stem walls beginning at main floor level;     Or     c) The same but with the air heaters higher up, in some of the south facing traingles;     Or     d) A solar attic, in this case with say 2 triangular glazings in the front of the top pentagon ( though I sense you mean more of the south roof area and therefore more vertical space for the attic  ) with a floor of collector area with water being pumped from, circulated through and drained down to the water tank(s). Any references to solar attic designs might help me comprehend this a bit better, though  the dome complicates the angles involved. 2) Concrete blocks  to do pretty much the same thing. Partly what confuses me is the heights you envision; with a 28 ft dome, the ideal height in this case would be 14′ plus 3-4′ for a stem wall, assuming the floor is at the bottom of the stem wall, giving 17′ total height ( 8′ for first floor, 8′ for 2nd, 1-2′ for attic ). Is that correct? Also, I haven’t added my thanks for any and all help offered, it is very generous of you. Let me know when I need to stop pestering you or start sending cheques, or checks. I’m also still mulling over adding the concentrator, if Duane C. Johnson wants to jump in… Alec

Response:

– Hide quoted text — Show quoted text ->     Or >     d) A solar attic, in this case with say 2 triangular glazings in the > front of the top pentagon ( though I sense you mean more of the south roof > area and therefore more vertical space for the attic  ) with a floor of > collector area with water being pumped from, circulated through and drained > down to the water tank(s). Any references to solar attic designs might help > me comprehend this a bit better, though  the dome complicates the angles > involved. > Partly what confuses me is the heights you envision; with a 28 ft dome, the > ideal height in this case would be 14′ plus 3-4′ for a stem wall, assuming > the floor is at the bottom of the stem wall, giving 17′ total height ( 8′ > for first floor, 8′ for 2nd, 1-2′ for attic ). Is that correct?

Just thought: are you envisioning the top pentagon-cupola "extrusion" that is popular on many dome designs? Alec

Response:

>[snip] >So the concrete column ( which I failed to realize…got stuck on the 4′ >dimension ) acts as a thermal "chimney". I’m starting to like the water >ideas anyhow. >[snip]

         In Canada isn’t it important to consider that the whole system may end up frozen some time, and water would make a mess after busting pipes? Joe Fischer

Response:

>[snip] >So the concrete column ( which I failed to realize…got stuck on the 4′ >dimension ) acts as a thermal "chimney". I’m starting to like the water >ideas anyhow. >[snip] >          In Canada isn’t it important to consider that the whole system > may end up frozen some time, and water would make a mess after > busting pipes? > Joe Fischer

Doesn’t water freeze elsewhere? The first floor ceiling would be ok, since it is heated by air from the air heaters in the walls- no water to freeze. From what I can tell, the solar attic having a draindown floor means the water would drain down at the end of the day. Thermosyphoning or a pump I should think, although I understood drain down systems to be problematic, at least in flat plate collector systems. Or, perhaps Nick means for it to be contained freely, like Thomason’s roof trough system, so freezing a small amount left in the system wouldn’t matter. Alec

Response:

>>You might store cloudy-day heat in a 70 ft^3 ~4′ cubical tank under the >floor that cools from 130-80 F over 5 cloudy days, and warm the dome on >a cloudy day by pumping some water up through the ceiling. The attic floor >needn’t contain much water, just enough for solar collection during the day. >It would have insulation underneath, and maybe a few transparent patches for >daylighting. With separate average and cloudy day tanks, you might avoid >the air heaters. >…I’d like to have a floor plan which only includes a semicircle for the >second floor…

In that case, 100 F air can’t pool below the second floor, unless it has a big lip, but you might store overnight heat below the attic floor… >1) One or Two separate, one-for-cloudy-day ( stored under the first >floor? ), water heat storage units in the form of 2 Thomason’s-pressurized- >tank-within-nonpressurized-box units being heated by:

I’d say two, or a stratified tank, with warmish water at a higher solar collection efficiency providing heat on average days and hotter water at a lower collection efficiency for cloudy days. You only need one pressurized tank to heat water for showers. >    d) A solar attic, in this case with say 2 triangular glazings in the >front of the top pentagon ( though I sense you mean more of the south roof >area and therefore more vertical space for the attic  ) with a floor of >collector area with water being pumped from, circulated through and drained >down to the water tank(s)…

And enough solar aperture to heat the dome, with 598 Btu/ft^2-day on a south wall and 369 on a horizontal surface. >Partly what confuses me is the heights you envision; with a 28 ft dome, the >ideal height in this case would be 14′ plus 3-4′ for a stem wall, assuming >the floor is at the bottom of the stem wall, giving 17′ total height ( 8′ >for first floor, 8′ for 2nd, 1-2′ for attic ). Is that correct?

Sounds OK to me. With a 4′ stemwall, the attic floor surface would be Pi(14^2-12^2) = 163 ft^2, no? Nick

Response:

>Just thought: are you envisioning the top pentagon-cupola "extrusion" that >is popular on many dome designs?

Not raised, but possibly transparent. Nick

Response:

>From what I can tell, the solar attic having a draindown floor means >the water would drain down at the end of the day.

Yes, it would drain back to the basement when the pump shuts off, without any solenoid valve that needs to open. >…I understood drain down systems to be problematic, at least in flat plate >collector systems.

Outdoor pipes can slump and trap water that freezes, but in this case, all the water would be inside. >…perhaps Nick means for it to be contained freely, like Thomason’s roof >trough system, so freezing a small amount left in the system wouldn’t matter.

Also true. The water might be in a shallow flat poly film duct or a shallow flat poly film pond with a poly film cover, with a low point in the center. Nick

Response:

>Partly what confuses me is the heights you envision; with a 28 ft dome, the >ideal height in this case would be 14′ plus 3-4′ for a stem wall, assuming >the floor is at the bottom of the stem wall, giving 17′ total height ( 8′ >for first floor, 8′ for 2nd, 1-2′ for attic ). Is that correct? > Sounds OK to me. With a 4′ stemwall, the attic floor surface would be > Pi(14^2-12^2) = 163 ft^2, no? > Nick

Since there is no alt.geometry.for.the.guy.who.fell.asleep.during.the.bit.on.spheres.and.inte r secting.planes, it took me time to get off my butt and let some blood into my brain. Envisioning the inverted right triangle that results from the height of the attic from center, and the radius of the (hemi)sphere finally did it. Now some more questions… Is 4′ as high as I can go with air heaters in the south wall, while keeping the system passive? Noting the chimney formula cfm = 16.6 Av sqrt(HdT), how do you figure out the size of the vents in the sunspace? Are they sized proportionally to the sunspace volume to facilitate a moving stream of air? Is it prudent to overbuild a solar system somewhat? In the dimensions given for the air heater the size was greater than the calculation ( unless it’s to account for any framing material that might block light ):      "…so the extra heat might come from 25K/332 = 78 ft^2 of air heaters, eg       a 4′x25′ skirt around a 4′ stemwall…", this being approximately 28% greater than 78 ft^2. Alec

Response:

Sorry, just thought of another question, since I can now figure out approximate floor area for upper floors…how should I calculate the window to floor area proportion? 96 ft^2 / 1363 ft^2 = 7%. I got the 1363 from all three floors. Alec

Response:

Ack! Wait! I’m finding some of the answers, as previously pointed out, at the archive…I’ll be back if I can’t find the answers there, sorry. Alec

Response:

>Is 4′ as high as I can go with air heaters in the south wall, while keeping >the system passive?

Maybe not. >Noting the chimney formula cfm = 16.6 Av sqrt(HdT), how do you figure out >the size of the vents in the sunspace? Are they sized proportionally to the >sunspace volume to facilitate a moving stream of air?

You might use desired sunspace air temps and the amount of glazing and the peak sun power. For instance, if 70 F living space air enters a sunspace on a clear 30 F day and 100 F sunspace air returns to the living space and a transpired mesh collector near the north wall keeps room air in the bulk of the sunspace and near the glazing and makes 100 F sun-warmed air to the north of it, and 250 Btu/h of sun falls on each square foot of R2 sunspace glazing with 80% solar transmission and 200 Btu/h enters the sunspace and the glazing loses (70-30)1ft^2/R2 = 20 Btu/h to the outdoors, a sunspace with 100 ft^2 of glazing would gain 18K Btu/h, net. With 100 F air behind the mesh and 70 F air on the other side of the living space wall, dT = 30 F and 18K Btu/h = 30cfm, since 1 Btu/h will heat a 1 cfm airstream about 1 degree F. If H = 8′ and 18K Btu/h = 30×16.6Avsqrt(8×30), Av = 2.3 ft^2. >Is it prudent to overbuild a solar system somewhat?

Sure. You might do this to better ensure that it works on partly-cloudy days or to make up for construction mishaps like insulation gaps or air leaks or to keep framing on 4′ dimensions. You might want to avoid cutting 4′x8′ sheets of plywood or foamboard or 4′x50′ rolls of polycarbonate plastic or 10′ or 20′ wide rolls of EPDM rubber. Nick

Response:

>…how should I calculate the window to floor area proportion?

The less the better, I’d say, unless you are bound by building codes, which talk about fire exits and 4-8% of the floorspace for ventilation. Windows are poor insulators, and expensive, compared to good walls. They can provide solar heat on an average day, but unlike air heaters, they lose heat at night and on cloudy days and increase the size of average and cloudy-day heat stores. One might exit through an insulated hatch, and there are other ways to provide ventilation. Nick

Response: