Iain McClatchie wrote:

Nick You might "open windows" with a thermostat and an exhaust fan...

If I'm right about that 800 ft^2 of glass, and if 400 ft^2 of it is
passing sunlight, then he's getting 332k BTU on those February days...

.... 400 ft^2 of single glass might pass 0.9x250x400 = 90K Btu/h in full sun.

While the sun is up, he'll lose 50k BTU through the windows.

.... and lose (70-56.5)400/R1 = 5.4K Btu/h.

Assuming he loses a good chunk elsewhere and has no significant
thermal storage...

.... Significant improvements are possible.

he'll need to vent over 1000 cfm at 80F (from outside 55F).

.... cfm = (90K-5.4K)/(80-55) = 3384, approximately. With your command
of numbers, you might consider a different preoccupation, eg surgery :-)

Not exactly an 8" NuTone bathroom fan.

Maybe 3 $12 20" window box fans, if he wanted to waste all that solar heat.
Or 2 $55 2470 cfm Lasko fans, or a single whole house fan.

Nick My advice: forswear thy fossil fuels and improve the solar
Nick performance. It's easy, with high-school physics and simple
Nick arithmetic.

...and many, many tons of water.

With enough insulation, it wouldn't need any water, nor windows.

A 32' R32 cube with no air infiltration would need 24h(65-48.7)5x32^2/32
= 62.6K Btu on an average Jan day in SF. It might come from 18.3 kWh/day
or 550 kWh/mo of internal electrical usage, 2/3 of the 833 US average.

With no internal heat gain, it might stay warm for 5 cloudy days with
5x62.6K/(130-80)/2000 = 3.13 tons of water cooling from 130 to 80 F in
a 3.13x2000/1024/62.33x12 = 1.2" layer under the 4th floor ceiling. If
A ft^2 of R1 sunspace glazing with 90% solar transmission over the R32
south wall gains 945A Btu/day and loses 6h(130-56.5)A = 441A and makes
62.6K Btu net, A = 124 ft^2, eg an 8'x16' sunspace or $200 patch of
polycarbonate solar siding.

Solar house heating is extremely easy in San Francisco.

Nick