You are viewing an older revision! See the latest version
Weather Station
The weather station team was tasked with recording all weather variables and sending the data to the interface. This team was also responsible with receiving time from the internet to synchronise with other machines.
Design Brief¶
Set up the sensors in appropriate circuits. Write the code allowing our mbed to receive useful data. Calibrate sensors which output only voltage readings so it can be interpreted as useful weather data. Find a reliable way to send weather data to the interface team so that they can read it and display it on their screen.
The Build¶
Temperature and Pressure Sensor¶
For this we used the ms5637 with data sheet : http://www.farnell.com/datasheets/1756129.pdf

For setting up the sensor we used the circuit diagram given on page 11.
By using the I2c interface for SDA /SCL and connecting VDD to 3v3 voltage supply on the mbed the sensor was relatively easy to setup.
For the code, multiple libraries can be found already on the mbed website and the code was very simple to setup just to print the data to the pc terminal.
LDR - Light Dependent Resistor¶
We would be using the LDR to measure light intensity in Wm^-2. Since the device acts as a variable resistor we set it up in a potential divider circuit using the " " voltage source on the mbed. Reading out a voltage was very simple, however the calibration required to gain accurate light intensity values as measured against a light meter proved to be very difficult. Since the LDR is made up of a semiconductor material it is also highly affecting by changes in temperature so truly accurate readings with this sensor would prove tricky.
Using the light meter we took values of light intensity for the voltage printed out by our device. By plotting the data using matlab we found that it's characteristic definitely could not be desribed as linear. As the graph resembled a 1/x function, we used the form y=constant/x .Using matlab's linear regression function we found a value for the constant. When the curve was displayed against the data we decided this to be a close enough fit.