Where should a temperature sensor be placed for most accurate results?

OLDmarty

Senior Member
Hi All,

As the title suggests, "where" is the best place to mount a temperature sensor which is of course connected to a picaxe project.

For outdoor use, i'm guessing it should NOT be in direct sunlight, so it must have a sun shield over it, but also can't be mounted in a shaded area because that might not be true temperature???

I'm also guessing it shouldn't be mounted directly on a roof surface or next to a wall that would retain heat, which would also give misleading readings (brick wall etc).

At this stage, i believe the sensor should be mounted on a small mast (maybe some TV antenna hardware) and of course a sun-shield over the top of the sensor.

Has anyone tried this before? and what methods did you use?

Thanks in advance.
Marty
 

premelec

Senior Member
it depends on what you want to measure - the temperature is different all around the place... if you want north house air temperature - or being in Australia south face air temperature [:)] - the most fun is had with lots of temperature sense units around the area to see just what the differences are...
 

MurrayJ

Senior Member
What you need is called a "Radiation Shield For Weather Station". Basically it blocks direct sunlight but allows airflow.
 

westaust55

Moderator
Assuming you want air temperature such as a weather station would provide, Suggestion for location is:
1. In a radiation enclosure / Stevenson’s box
2. 1.5 meters above the ground
3. Over a grassed area (avoid paved or other sealed areas
4. Away from trees to prevent shading effects to solar radiation and wind flow
 

SteveDee

Senior Member
You can really only take a reference reading because "the temperature" in the room of a house or in your back garden varies so much from point to point.

The temperature gradient in a room in your house is particularly interesting. We are often given "advice" this time of year (it was -5'C in sunny southern England the night before last) that your heating thermostat should be set to (say) 20'C as that's plenty high enough. But when you are sitting in your favourite chair, the temperature around your head may be ok but your feet may be frozen to the floor! So an intelligent room thermostat probably needs multiple sensor positioned around the room and at varying heights.

Stevenson's screen: https://en.wikipedia.org/wiki/Stevenson_screen (this would look interesting positioned in the middle of my lounge!)
 

hippy

Technical Support
Staff member
My idea for doing it easily and cheaply would be to mount the sensor on the top of a hollow pole. That allows mounting at height and cables can come out of side or bottom.

Put a section of white PVC pipe around that. Use the type which can have end caps fitted and that turns it into a can. Drill holes around the top and bottom edges and also in the bottom. Add some deflector flanges to keep the rain out of the holes.

Code:
  .-----------.
 /             \ 
/               \
  |    ___    |
  |   |___|   |
  |    | |    |
 /     | |     \
/     /| |\     \
  `- --| |-- -'
      \| |/
       | |
You could probably mount an umbrella style cone over the top for shading from sunlight, possibly taking the mounting pipe out of the top.

Have a look at the page below for some ideas. My idea above seems to be a simplified version of the very bottom picture. Given a number there don't have flanges over the holes they might not be needed except for at the top and a covering cone may do.

http://www.weatherforschools.me.uk/html/weatherboxes.html

A nice DIY design using standard louvre vents here ...

http://www.instructables.com/id/Stevenson-Screen-weather-station
 

AllyCat

Senior Member
Hi,

so it must have a sun shield over it, but also can't be mounted in a shaded area because that might not be true temperature???
It depends how you define "true temperature", but as far as a "Weather Station" is concerned, WA55 gives the exact specification in post #4.

But a real "Stevenson Screen" is quite complex (and large), so some of the better weather stations (e.g. Davis FARS) use a fan to circulate air past the temperature sensor. Powering that from a PV (solar) panel is quite a good solution, because you mainly need the fan to run when the sun is shining! ;)

To give an indication of how "difficult" it can be, the best-known "budget" weather stations use this solar screen (note the main photo is upside down!), but when it's in direct sunlight, I've measured a rise in temperature of more than 5 degrees C above the "true" air temperature, even in UK. Hence I mounted mine "in the shade".

Cheers, Alan.
 

techElder

Well-known member
Find out what our current crop of "scientists" do to position temperature measuring equipment and do the opposite!
 

premelec

Senior Member
FWIW an interesting temperature measurement is looking straight up with an IR thermometer - that gives you information on cloud, haze etc... often times below freezing on clear day... and preparing you for your balloon ascent ;-0
 

fernando_g

Senior Member
I once performed an experiment, with two temperature sensors in almost identical circumstances, but one had reflections from a concrete pavement, and the other was in the grass.

The temperature difference at noon on a sunny summer day, was about 4 Celsius higher, even though the sensor was at least 3 meters away from the driveway.
IR reflection is very real!
 

OLDmarty

Senior Member
Thanks everyone for your input.

Great to see i was exactly on track with all the conditions to be weary of.

I, in fact, invented my own Stevenson Shiled (box) without even knowing what it was called beforehand lol.
 

erco

Senior Member
Marty: I'm pretty sure you want to use an array of Sensirion sensors in Stevenson boxes in Faraday cages at the Lagrange Point. :)
 
Top