One person.
It seems like you've supported my point, not refuted it.
But I appreciate your analysis, even if I dispute your conclusion! Thanks!
However, I made an error! It turns out I left out the coefficient of performance in my calculations. So it's actually 105 watts, 8 m², and about US$1000:
You have: (8 liters / day) * (41 kJ / mol) / (18 g / mol) * (1 g / cc)
You want: W
* 210.90535
/ 0.0047414634
You have: (8 liters / day) * (41 kJ / mol) / (18 g / mol) * (1 g / cc) / 2
You want: W
* 105.45267
/ 0.0094829268
You have: (8 liters / day) * (41 kJ / mol) / (18 g / mol) * (1 g / cc) / 2 / (13 W / m^2)
You want: m^2
* 8.1117442
/ 0.12327805
My guess of US$1000 for 8 m² is based on my guess that 8 m² is about 8000 watts of solar constant, and at 16% efficiency that's 1280 peak watts, and people quote PV modules as below US$1/watt, which I assume is a peak watt. The actual price right now is about US$0.55/watt, depending on where you are, so a more precise estimate is actually US$700: You have: 8000 watts * 16% * $ 0.55/watt
You want:
Definition: 704 US$
My friend who's getting those 13 W/m² has her panels in England, though. So that's kind of a worst-case number. You'd need a fraction of that (like, say, 20% or so) in an equatorial desert.If PV modules return to their previous price trend (they were dropping fairly smoothly by 36% per year in 2010-2013, but have been nearly plateaued since hitting grid parity in much of the world in 2013 and therefore having a hard time with capacity keeping up with demand) then they'd drop by another factor of 4. The Credit Suisse doc posted earlier today to HN https://doc.research-and-analytics.csfb.com/docView?sourceid... about how we are now in a "solar manufacturing oversupply" situation suggests that something like that might actually happen. Like what happened to DRAM prices in 1996 after the price-fixing cartel collapsed and they went from $40 a meg to $10 a meg almost overnight.
The pump were using right now consumes 104 W intermittently. For our next prototype, we're going to use a condenser unit from a walk-in freezer (1168W), a split A/C unit to pretreat the air (560W) MinnowBoardMax (15W), duct fan (208W), etc.
Typical coefficients of performance for air-source heat pumps, like air conditioners or dehumidifiers, are around 2; that means for every 2 joules of heat that it removes from the cold reservoir, it consumes a joule of electrical energy (dumping it as heat into the hot reservoir along with the heat it's pumping). The enthalpy of vaporization of water, at almost 41kJ/mol, is huge compared to the specific heat of air or even water itself, so that's a good approximation to how much heat you need to extract to condense the water. Water's molar mass is about 18 g/mol (two hydrogens of 1.0001 or so and an oxygen of 15.9994) so that works out to about 2.3kJ/g. At 10¢/kWh, this means your water costs 6.3¢/ℓ or US$78000/acrefoot.
This is a lot cheaper than bottled water, and a lot more expensive than what farmers can afford to pay for irrigation water, and a lot more expensive than reverse osmosis desalination, which I think is around $8000/acrefoot. (Corrections welcome!)
If we figure that a person needs 8 ℓ / day for cooking, drinking, and bathing (more or less what we use at Burning Man; if you're somewhere less hot and druggy then you might need less) which is about 210 watts. If a photovoltaic panel produces a 24-hour average of 13 W/m², which is what a friend of mine in England is getting, you need about 16 square meters of photovoltaic panels to power the device, which costs about US$2000 at this point. That's well within the bounds of feasibility, although it's a big enough asset to be a temptation to thieves.
So when you say, "I feel like I'm missing something here," I agree: you were missing quantitative understanding of the subject, and as a result your quantitative conclusions were completely wrong, even though they were based on correct qualitative understanding.