Make your own free website on


Night Sky Cooling
(or Lunar Cooler?)

If we are to make any major reduction in energy use, it will be in heating and cooling homes before it will be in transportation. I refer to it as lunar cooling because it is almost the opposite of solar heating.

Night sky cooling is not a new idea by any means. There are accounts of ancient Egypt making ice using night sky cooling. Shallow clay dishes were placed on straw in an underground room that was open to the night sky. I think it was a combination of evaporation of water from the outside of the porous clay dish and the night sky that would make a thin layer of ice available by morning. It is why we find frost on cars in the morning when the temperature has not hit freezing.

How much cooling is available?
Some references on the 'net give the night sky constant as 1/20 of the solar constant which would be ~1kw/meterē/20 or ~50 watts/meterē. Others give 5-8 watts per meterē. The difference is an order of magnitude. These are a few of the factors that affect it... dust in the air, thin clouds, thermal pollution from factories and homes, and possibly humidity. This is fortunate, because it keeps night time temperatures from being what it is on the Moon.

So far my personal best has been a 14.6 °F drop below ambient in the daytime!

1 watt is 3.412 Btu/h
1Btu/h is .2931 watt

We find
@ 1/20 solar constant = 170.6 Btu/h/meterē (144 Btu = melting 1 pound of ice)
@ 5-8 watts/meterē = 17=27 Btu/h/meterē
The usable cooling power should be somewhere in between. We will see in the following.

Where to aim it?
This is what I use to determine the optimum area of the night sky to point the parabolic. It is simply an infrared photo transistor (T 1 3/4 case, Radio Shack 276-145a)  epoxied into a brass or aluminum tube from the hobby store. A digital volt/ohm meter registers the relative infrared intensity. It may be best to insulate the tube, or at least wear gloves, but so far it suffices as is. Diameter of the photo sensor in .200 inches.

The night sky is quite noisy in the infrared. The transistor conducts when it receives infrared, so we want an area of sky that gives the highest resistance on the ohmmeter.

Milkyway View of Milky Way galaxy in infrared. (NASA).
It may be best to void letting the collector 'see' the Milky Way galaxy. Note that in winter the Milky Way is an East to West band that is farther South in the sky than it is in summer if you are in the northern hemisphere.

What type of collector?
My original plan for a cooler was something that resembles a photo multiplier tube, but working in reverse. It doesn't work because each element in the multiplier picks up heat from the ambient air, warming it. It might work in a vacuum, but that is beyond what I am willing to experiment with.

This begins my third year of experimenting with night cooling. So far, the drawing above right has given me the best results. An earlier design had HDPE plastic (almost transparent to infrared) over the parabolic. It was found that the ambient temperature warms the plastic, which in turn prevents maximum cooling. Air flows through the PVC pipe into the volume to be cooled. When I find what I did with the original graphs it will be posted here.

This is cheap, one moving part (the fan) simple electronic circuit most anyone can make.
Differential temperature switchThe used computer fan draws about 1.5 watts.
The circuit is so sensitive that holding my finger near it will make it switch. The same circuit can be used to direct air from a darkened box during daytime in winter for a boost in heating. (not tried yet)

A pair of remote thermometers are on the way so I can get some more reliable data.

5 April 09
The remote thermometers have arrived. I spend two days testing them for accuracy at various temperatures. The set up was placed outside late afternoon. The V shape configuration was used.

Points of interest are 7PM when the tube went into the shade, and 7:30 when the Sun went down. The area where there is significant cooling though it was still light out is very promising because dust in the upper atmosphere is still being heated by sunlight.

Useful cooling temperature may occur at 5:30 (daylight saving time), important in places like here in the Southwest.

night sky cooling
7pm is the best reading @14.1°F drop below ambient.

The spread increases again when the wind became calm. It will tested again with HDPE covering to prevent warm air from the wind warming the tube. Compare the wind in the table below to significant deviations in the graph.
             Wind        RH *
5:00pm  10-15     19%  tube in sunlight, high thin haze
5:30       5-10       17%      "
6:00       7            17%      "
6:30       7            18%      "
7:00       5            18%  tube partly in sunlight.
7:30       5            18%  dusk
8:00       5            24%  dark
8:30       5            24%
9:00       5            24%
9:30       3            22%
10:00     3            22%
10:30     calm
* Wind and relative humidity measurements were taken from the National Weather Service.

A few other factors surfaced... Heat from the brick wall of the house is picked up by the reflector. Ideally, it should be away from all objects. I was a little surprised that the yard light (spotlight type) I turned on to read the temperature also affects it if left on a few minutes (about .4 degrees). And finally, passage of the Moon over the collector (note graph above). 

NOTE: Three other thermometers read 43-44 degrees at 11:30 while the cooler read 25.0 and 32.4 ambient! This is because the others are near the house. A good example of how much heat a house can leak.

6 April 09
Same test, but with the tube spray painted black. Weather is about the same as last night.
Grid 3
I am confident now that relative humidity plays a major part. This is the best performance yet, 14.6° drop @ 7PM, and 14.5° @ 7:30.

7 April 09
Today I tested a film of HDPE to protect the tube from the wind. The graph is very much like above, but the differential slightly less even though humidity was down. The ability to cool in daytime is encouraging.

8 April 09
Overcast and low 80 temps did what I expected... tube temp a few degrees above ambient. Tried it both with and without the HDPE film. Next will be to design an intermediate between the crude setup now and a final version. It will have an adjustable V and the tube depth can be adjusted. Pivots will allow it to be aimed to an area of sky conveniently. I want to determine the optimum ratio of tube diameter to reflector size.

24 April 09
It was a bit windy, but I wanted to place a pound (16 ounces) of water in the solar cooker funnel and measure the temperature overnight. The cone will give a narrower portion of the sky.

         water  Amb    dif      RH     wind
2 pm   73      85.3  12.4  12%    23 mph  overcast, sun within cone view 
2:30    74.5   84.7  10.2
3:00  wind blew it over
6:30    76.5   86.9  10.4  12%      7 mph
7:00    75.6   84.6   9      11%      8 mph
7:30    71.8   77.7   6.9
8:00    68.2   4.1     5.9   8%      11 mph 
8:30    65.7   71.2   5.5
9:00    63.9   71.1   7.2   31%    17 mph clear sky
9:30    61.9   67.5   5.6
10:00  60.4   66.0   5.6   39%    12 mph
10:30  58.5   64.6   6.1
11:00  56.8   61.7   4.9   46%      9 mph
11:30  55.2   61.2   6.0  
12:00  54.1   60.0   5.7   43%    10 mph
12:30  53.1   60.1   7.0
1:00    52.2   59.4   7.2

low     34.8  42.8     8.0   HDPE bag had blown off

I need to do this again with the bag secured better. Then try it with 2 pounds of water to give a rough idea of how much water can be cooled for a given reflector size.

Some more recent data

19 Jan, 2012  
  I got one of those hand-held remote reading thermometers yesterday and went outside into clear area. 
Ambient was 32.0°F humidity 26%. Directly overhead seemed best with a reading of -1 to -1.7. Interestingly, it dropped to -5 just before a big noisy gust of wind. Away from overhead, temperatures read as high as +8.

  61°F,  22%, solar radiation 205 watts/mē.   The lowest reading was again directly overhead at 1.1 degrees. Highest (while not aiming it at the Sun) was 11°F. I seems that for this time of the year at least, aiming the cooler straight up and leaving it there is good.
NEXT I would like to see what my sky map will resemble when it is hazy, cloudy, and raining if we ever get rain.

24 Jan, 2012
11AM Overcast. Sky temp readings 19.1 to 29 degrees, with the best at about 45 degree angle toward the West.

Btu - Brittish thernal unit  A measurement for heat (NOT temperature)
HDPE - High Density PolyEthylene  The plastic film used for baking a turkey in a bag. Tranparent to infrared.
PVC - PolyVinyl Chloride  common in the home for plumbing pipes.
RH - Relative Humidity

Article Details > 3) Night-Sky Cooling
Night Sky Radiant Cooling System Saves Energy in a Commercial Building
Passive Solar Heating and Cooling
Ron Swelter's Tropical Cool Blog: How cold that night sky
How to Make and Use the Solar Funnel Cooker Cooling is near the end
Re: Is there refrigeration in ancient times

NOTE I am still very much interested in this, and experiments continue.

Copyright Đ 2009-2012
Contents of this site are the property of Bill Weller and may not be reproduced in any form without permission.
Copying HTML or other intellectual property violates Copyright Laws and bears serious penalties.