Skip to content
DIY Solar Book, home

Part VI: Staying CoolChapter 10. Passive and Low-Energy Cooling

Radiant Night-Sky Cooling Experiments

Surfaces that lose heat to a clear night sky, treated as experiments with measured limits.

Difficulty
Advanced
Cost
Varies
published

On a clear night a surface can send more infrared to the sky than the sky sends back. The surface cools. Dew and frost are the ordinary proof: the ground, a car roof, and a leaf often sit colder than the air. Clouds and humid air fill the sky with infrared and the effect shrinks toward nothing. Wind, once the surface is colder than the air, brings heat back. Still, clear, dry nights are the only nights the experiment is interesting.

Treat it as an experiment, not as a plan to skip air conditioning. The useful cooling is on the order of tens of watts per square meter under a good sky, and far less under a poor one. A few square meters will not carry a house. Shade, insulation, and night air when it is actually cool move more heat with less invention.

Difficulty

Level
Advanced
Time
A few clear nights with a thermometer, on a panel you can reach from the ground
Skills
Measuring a small effect and not promoting it into a roof pond

Cost estimate

Varies

A scrap panel, a thermometer, and a sheet of thin polyethylene can be a small experiment. A roof pond, a pumped radiator, or a commercial radiative film is a different budget and a structural question. Do not start on the roof.

Materials

ItemQtyNotes
A small dark panel on supports1Insulated on the back so the ground is not the thing you are measuring. Reachable without a ladder.
Two thermometers1 pairOne on the panel, one in the air nearby but shaded from the panel. An infrared gun on shiny metal lies.
A clear night and a cloudy nightBothThe cloudy night is the control. If the panel does the same thing, you are not seeing the sky.
Optional thin polyethylene cover1It slows the wind and still lets much of the infrared through. It is not a roof. It shreds in the sun. Take it down.

What you are measuring

Set the panel out after sunset, back insulated, face toward the sky. Read the panel and the air every little while. On a dry clear night the panel should fall below the air temperature. Dew on it means it crossed the dew point, which is evidence, and also a film of water that changes the surface. On a cloudy night the gap should shrink. If it does not, the panel is coupled to something else: a cold ground, a breeze off a lake, or a bad thermometer.

Write the sky condition down. "It got cool" is not a result. The air got cool too. The result is the difference.

The numbers stay small. Ten square meters at 50 watts each is 500 watts, about 1,700 BTU per hour, and only while the sky cooperates. A closed-up house on a hot evening can want more than that, and the hour before sunset is when you needed the cooling, which is when the sky is not yet a sink. The mismatch is the reason this stays an experiment.

Commercial films that reflect sunlight and still emit infrared are products with data sheets. A can of paint is not those films. Do not coat a roof and call it the same thing.

A clear sky can cool a surfaceA small panel faces a clear night sky and is cooler than the air. A cloud is shown cutting that path off. Dew is noted on the panel.Cloud cuts itPanel cooler than the airDew means it worked a littleTry this on a small panel. You still need shade.
A small panel looks at a clear sky and runs cooler than the air. Clouds cut the path. Dew means the surface crossed the dew point.

What not to build because the panel worked

A roof pond is a historical cooling idea: water on a roof, uncovered to the night sky, covered in the day. It is also a structural load, a leak above every room you own, a mosquito source, and a fall hazard at the edge. This page does not tell you how to build one. If the small panel convinced you the sky is cold, the next step is still shade and a night flush, not a pond.

A pump circulating water through a radiator aimed at the sky adds electricity, freeze risk, and a leak. It can be a legitimate research setup. It is a poor first project, and it does not belong on a roof you visit at night. Freeze will split the collector the first cold clear night, which is exactly the night the radiation is strongest.

Do the experiment at ground level. Log three clear nights and one cloudy night. If the difference is a degree or two, you have measured the effect and you can stop. If you want the house cooler tomorrow, close the west shades in the morning.

  1. Build a small panel you can read from the ground

    Insulate the back. No ladder. No modification of the house roof.

  2. Compare panel temperature to nearby air

    Clear night and cloudy night. Write down wind and dew. Distrust a single infrared reading on bare metal.

  3. Convert a hopeful number into watts before you get excited

    Tens of watts per square meter is the right neighborhood on a good night. Compare that with the size of the room.

  4. Take the plastic down

    Polyethylene is an experimental windshield, not a building cover. It fails in sunlight.

  5. Cool the house with the tools that move more heat

    Shade, a tighter envelope, night air when it is cool and dry, and mechanical cooling when the weather is dangerous.

Further reading

These links go to outside sources. This book cites them by link and does not copy their words or pictures.

General education, not a permit or a stamped design. Follow the safety note in the guide, read the safety disclaimer, and hire a licensed trade when the work requires one.