I bought an Oregon Scientific WMR-968 weather station in 2002. It is a wireless unit, designed to be mounted outdoors, and has solar panels to recharge the batteries in each unit.

However, it lacks any sort of radiation shield/Stevenson Screen for the temperature/humidity sensor. Since exposure to sunlight causes errors in temperature readings the sensor must be carefully placed and out of direct sunlight. Radiation shields address this problem, protecting the sensor from erroneous readings due to radiation from the sun, both direct and reflected.

In lieu of a radiation shield the poor man’s/easy solution is to mount the temperature sensor under the eaves of one’s house. However, this will inevitably result in erroneous readings due to the radiated heat from the house itself as well as from the ground and any other nearby objects. On hot days I have measured a minimum of three to four degrees increased temperature (temperature error) from a sensor mounted bare under the eaves of a house and one in a functioning radiation shield.

Therefore the temperature sensor needs to be mounted in a radiation shield or Stevenson Screen to operate most accurately. There are many manufacturers of radiation shields and Stevenson Screens. However, commercial variations of radiation shields are specialty items and therefore tend to be expensive.

Passive shields tend to be priced at $35 for hobby variants and up, while commercial variants can be $150 or more.

The Stevenson Screen is a box with double-louvered sides that encloses the temperature sensor. Multi-plate radiation shields (or Gill Radiation shields) are a series of shaped plates stacked vertically with space between each plate and a hollow core in which the temperature sensor is placed. Both are usually painted white to reflect sunlight and are intended to enclose the temperature sensor to prevent having solar radiation directly affect the temperature sensor while at the same time allowing air to flow around it.

Since the multi-plate radiation shield design is easier to build and more compact I opted to construct that variant. There are two sub-types of radiation shields: passive (naturally aspirated/ventilated) and active (forced aspirated/ventilated). Passive radiation shields just offer protection from the direct sunlight/radiation and inevitably have some temperature errors due to heating of the shield itself in bright sunlight. Active radiation shields are passive shields with fans to move air through the shields at times of peak radiational heating creating lower temperature errors.

Here is a comparison of Davis Weather active and passive shields.

Not having disassembled a commercial radiation shield I made my best guesses at how they were built, aided by several other websites documenting homemade shields.

My first radiation shield was made from plastic cereal bowls and threaded rods. UV light and the weather will cause plastic to dry out and degrade over time, so painting is a requirement if you choose to use plastic outside.

Initially it was a passive shield which seemed to work fine until the warmer months, so I added a solar powered fan to actively aspirate the unit. I discovered that using a solar panel to power the fan worked optimally, as it runs the fan faster as the sunlight increases thus countering the increased solar heating.

I recently decided to rebuild the radiation shield using metal cake pans.

Up to that time I had always left the bottom of the shield open, believing that was necessary for good airflow through the shield. However after trial and error experimenting with my shield this year I discovered that on very hot days reflected radiation from the ground was affecting my temperature readings.

Therefore I added an insulated bottom to the radiation shield.

The radiation shield was constructed from nine 8” diameter round cake pans, 5/16” threaded rod, ½” O.D. poly tubing, and rigid foam insulation.

The pans are connected using three sections of threaded rod, using 1.25” long ½” O.D. polyethelene tubing as spacers. There are three holes in each pan for each threaded rod.

Five of the pans have 3 ½” holes cut in the center of them to house the temperature sensor. Two of the remaining four pans have a circle of 1” rigid foam insulation glued to the inside.

This is the 8 inch diameter cake pan that comprises the plates of the shield		The first two sections of the shield before the second section has been seated. Note the rigid foam on the inside of the top plate.		The first two sections of the shield. Note the plastic tube spacers on the threaded rods.
The top section of the shield		Looking down the core to see the fan.		The complete top section of the shield before addition of the studs for the bottom section.
The top section of the shield with the bolts for the bottom section.		The shield mounted outside. The silver/unpainted pans are the bottom section of the shield.		The weather station. The solar panel that drives the fan is mounted on the wood post above the shield.

The pans are assembled with an insulated pan on top, a pan without a hole in it, then four pans with 3 ½” holes in them, the temperature sensor bracket (with the temperature sensor attached to it), one pan with a 3 ½” hole in it, then another insulated pan, and then a pan without a hole in it on the bottom.

The topmost pan with a 3 ½” hole has the 80mm 5 VDC fan mounted over the hole exhausting air downwards (I tried drawing the air up through the shield but found that blowing the air down worked better).

The shield is built in two sections. The top section includes six pans, including the temperature sensor itself. The bottommost pan in this section has another set of three threaded rods to attach the remaining pans after the temperature sensor is mounted.

I’m working on a more elegant solution for attaching the lower assembly, as mounting the current design is a bit troublesome.