Systems map::

I have three different ways to measure and observe weather, the newest of which is my fully automated electronic sensor array.

My sensor array design project:

The sensor array which I have helped design and built (thanks daddy!) measures, Light levels, Relative Humidity and the Temperature.

I already have a commercially built system to measure temperature and humidity, however that is getting old and we have noticed some errors in the data occasionally, however fundamentally the old system relies on manual input to my blog. My new system automatically collects data and graphs it.

I also included a new type of sensor, the light sensor, this is because I wanted to compare light levels to temperature and see if there was any correlation.

Sensors in detail:

Temperature: The electronic temperature sensor I am using is a TMP36. This is a simple device that outputs a fixed voltage relative to the temperature. It can operate in the range of -40ºC to 150ºC, so it will easily be able to measure even the most extremes of weather we are likely ever to have.

The temperature sensor is housed in a Stevenson screen this is so that direct solar heating from the sun, or wind-chill doesn’t affect the measurement, air can gently ventilate through vents giving the ambient temperature.

Humidity:The humidity sensor I am using is a Humirel  HF3223, it is a self-contained module that outputs a variable frequency square wave in the range of 8-10KHz relative to the measured humidity.
We apply a simple formulaRH = (9740/Freq)/18 to convert the frequency to humidity (RH).

Like my temperature sensor (visible at the bottom of this picture) the humidity sensor is also housed in the Stevenson screen, so that wind chill, and rain resting on the sensor don’t affect the readings.

Light: The sensor I am using to measure lights levels is a simple Light Dependent Resistor (LDR). The LDR’s resistances changes with the amount of light. The brighter it is the lower the resistance.
This variable resistance is applied to a fixed reference voltage and the resulting changes are measured as an analogue voltage input.

The light sensor is housed in a radome and is placed on top of the mounting box so that it isn’t shadowed by anything above it.

How it all Communicates:

Once a minute a cronscript running on Meteo (our weather systems server) sends a http request to the sensor array (actually an aurdino microcontoller development platform) via Ethernet networking. The microcontroller polls the individual sensors and pull back data from each, this data is then sent back via a http response as a simple comma delimited list of values. This is  interpreted by the cronscript and saved to a database.

This data from the database is the used to produce graphs for each, day, week, month and year and a “unique colour” for the moment are creating using other scripts

(A full explanation of how this all works, withy diagrams and source code will be avilable as a pdf here shortly.)

I design and built this system because of limitations of the equipment in my other commercially produced weather station.

Commercial weather station:

My maplins weather station equipment was installed about three and a half years ago and consists of the remote mounted sensors – (Wind, Rain, Temperature

and Humidity), and the internal display unit which also collects barometric pressure.

The external sensors are mounted on a tubular mast approximately 3 meters above ground level in a reasonably unobstructed corner of my garden a short distance from the house.

I take the majority of my daily readings direct from the display unit, but some measurements (such as cloud cover) still require a trip outdoors.

The external sensors:

Rain gauge: My weather station uses a fairly ordinary tipping bucket type rain gauge. As you can see from the image it has a collection area where rainfall lands, and is then funneled down into the measurement system inside.

This is similar in operation to a see-saw, or  a pair of scales. Basically it has two buckets one of which is always under the collection funnel. The rain water drains into a bucket until a certain amount of water is in the bucket, when its weight reaches a critical point, forces the balance arm to tilt, and this pours out the water – thus also moving the other bucket into place ready to fill again. Each time this tilting action takes place a magnetic sensor switch is closed and the count of flips is increased by one. (Each flip equates to about 0.03mm of rainfall). The total number of flips during any single counting period gives the total rainfall for that period. (48 seconds) This data along with data from all the other sensors is sent wirelessly back to the display unit indoors.

Originally the rain gauge was mounted at the top of the mast with other sensors, but it was noted that we would occasionally see false readings of rainfall when no rain had actually fallen. This was tracked down to vibration from the wind at the top of the mast causing the balance arm to occasionally flip of its own accord. After some thought, I decided it would be better to move the rain gauge to a point much lower down the mast, just above its rigid ground post. This move seems to have been effective as since then I have not noticed any false readings on the rainfall sensor.

Temperature and humidity sensor:
The external temperature and humidity sensors (along with the power supply and radio transmitter) are housed inside a the units “Stevenson Screen” (named after Thomas Stevenson who first made one in 1864).

The main purpose of a Stevenson Screen to prevent direct sun light heating the sensors directly rather than the sensors reporting air temperature etc – thus giving rise to false readings. The shape of the screen also helps prevent precipitation (rain, snow and sleet) from
entering the sensor housing and causing false readings too.

Anemometer: The anemometer on my weather station is one of the typical “three cup”
designs measure wind speed by counting revolutions / time. The faster the tree cups spin the higher the wind speed.

It is mounted highest up on the pole so it is clear of any other objects that would obstruct the wind.

The weather station can monitor speeds in meters per second, kilometers per hour, miles per hour or Beaufort scale. (My weather logging is usually done in kmh)

My Current weather station does not have a system for measuring wind direction electronically. I am currently looking at ways of adding this feature as a future electronics project, but for now my direction measurements are taken as a direct manual observation.

The internal systems::

Display: The display system forms the main part of my weather monitoring system.  It receives data every 48 seconds from the equipment outside, calculates values and displays results on its LCD (liquid crystal display).
The unit also contains the pressure gauge (as previously described), and an internal temperature and relative humidity sensor. The pressure gauge is corrected to mean sea level values, as my house is at about 137m above sea level.

As well as current readings it has features to display trends in data, history over the last 24 hours and minimum and maximum recorded values. It uses some (very!) basic heuristics to offer a short term graphical forecast of expected weather conditions.

My weather book: My weather book  – published by the Royal Meteorological Society – is probably one of my most important tools. I have used it to record daily weather data every day for three years. I can record wind speed and direction, rainfall, cloud cover in eighths, temperature and other observations such as “big cumulonimbus cloud over head.”  This data is entered into my weather computer at a convenient time and uploaded to the weather blog automatically. It also forms the basis of my weather archive which allows me to compare how conditions change from day today, week to week, and year to year.

NEW – Satellite system: I have recently added a purpose built weather satellite receiving system which allows me to capture and process live weather satellite images direct from the NOAA polar orbiting satellites. This has been integrated into the weather system computer. Live pictures from several daily passes can be found in the Sat Images section of the site.


%d bloggers like this: