Since my last blog post, many people have asked me “Michael, how is your weather station doing?“.
And the answer is ‘not so well’: almost every element of the station has a problem.
- The anemometer has stopped rotating.
- The wind direction indicator is stiff
- The rain gauge appears to always read zero but I am not sure if it is actually broken
- The thermometer screen is ineffective – allowing the thermometer to warm rapidly by two or three degrees in sunlight.
- The hygrometer mis-read periods when we have 100% relative humidity
- And the barometer is slightly wrong
So only the thermometer appears to working correctly! But despite that, I still consider the station to have been good value, and I am taking steps to rectify each one of these faults.
In fact, many aspects of the design appear to be excellent and with a little bit of tweaking I am hoping for many years of service.
The problem with the anemometer is that the bearing on which it rotates has seized up. The result is that the anemometer simply doesn’t rotate no matter what the wind speed.
Aside from the low quality bearings, the design is otherwise admirable. Voltage pulses are generated each time a magnet in the rotating head moves past a reed-relay in the housing. This should be highly weather-proof and have a very long service life.
Although it sounds complicated, replacing the bearing is not so hard and I have ordered a new one. To do this one just needs to specify the inner diameter of the bearing (that sticks on the weather station) and the outer diameter (that sticks to the anemometer head). So I hope that by next week I will have the anemometer up and working. I also ordered a second bearing because the wind direction indicator also felt a bit stiff.
The rain gauge appears to be working when I test it, but on some days it doesn’t seem to register any rain.
I took apart the gauge and I was pleasantly surprised at the quality of construction. I couldn’t see anything that would break in service – but I still haven’t understood why it isn’t working!
I may have to change either the magnet or the reed relay.
The thermometer screen
I noticed there was a problem with the thermometer screen by looking at the Met Office’s Weather Observations Website. There I could see that when it was sunny, my thermometer read up to 3 °C hotter than those in weather stations within a kilometre or so.
I made a couple of experiments with thermometer screens and eventually constructed one out of disposable plastic bowls (See picture at the top)
The bowls were thin and slightly transparent, and so I put a layer of aluminium foil between two bowls to make an object which would allow air to flow to the temperature sensor, and yet shield it from direct sunlight.
By the time I tested this arrangement, sunny days were becoming rare, but it did seem to reduce this rapid heating. However, I will have to wait for spring to test it fully.
Hopefully by then I will have engineered a more robust solution that doesn’t look so awful!
Being married to a world authority on humidity measurement I knew there would be problems with the hygrometer.
The basic behaviour of the sensor seemed approximately correct, but I noticed that it never read 100% humidity.
I took advantage of recent foggy weather to adjust the Cumulus software relative humidity reading to 100% – the stable temperature, absence of direct sun and the presence of billions of droplets of water in the fog ensure the air really is fully saturated.
Living near NPL, I have the great convenience of being able to check atmospheric pressure using NPL’s online barometer.
Anyone can use this, but the further away from NPL one goes, the less relevant the reading becomes. However, if one waits for a period of high pressure, with low winds and clear skies, the readings are probably reasonably accurate within 50 km of London
Using this reading – and adjusting for the difference in height above sea level, I made a small adjustment in the Cumulus Software to get the correct reading of atmospheric pressure within about 1 hPa (hectopascal).
Despite all these problems, I still think the weather station (at £90) represents exceptional value. It’s basic measurement of temperature seems to be correct, and the solar-powered logging and wireless connectivity is fantastic!
Making a significant improvement would involve purchasing a station that would cost £500 to £1000, and these stations are not perfect either.
So given the poor measurement site, it doesn’t seem worth spending that much money. However spending a few pounds improving this site seems like good value.
One of the reasons for the lack of recent articles is that – aside from being crazily pressured at work – I am attending a FutureLearn course on programming (in Python) for the purpose of analysing data. My hope is to automate the monthly analysis of data, which at the moment is proving very time-consuming.
With a trailing wind (which I should be able to measure!) I hope to be able to show some of my data here next week.
What interests me is that I need to do exactly the same things that climatologists do the data in order to extract trends. But actually doing something oneself, is difficult, but highly instructive.