|home page > weather station||Data: 24 hours | 7 days | 28 days | every year||site info / site map|
|rain sensor||EasyWeather data file format||weather station memory map|
In April 2008 I bought an electronic weather station. It's branded "Elecsa AstroTouch 6975", but a brief web search showed that it's functionally identical to several other brands, such as Watson W-8681 or WH-1080PC / WH1080 (the WH1081 is similar, but doesn't have a radio controlled clock). One thing that sets these apart from other weather stations is that they have a USB socket to connect to a computer. My station came with some Windows software called EasyWeather, but I've connected it to an embedded Linux box (an Asus WL-500gP V1 router) so I can upload "live" weather data to this web site.
The first two pictures show my first installation of the weather station's outdoor sensors, mounted on my garden fence with a TV aerial standoff bracket. At the top are the wind speed and direction sensors. These are far too near the ground (and other obstructions) to give meaningful measurements, but I should at least be able to judge relative wind speed from day to day. Half way down is the rain sensor. Last of all is the temperature & humidity sensor, which also contains the radio transmitter. Despite having a white louvred enclosure this is badly affected by strong sunlight, over-reading by 5 centigrade degrees or more. I need to make some sort of shield for it, perhaps from a length of plastic pipe. Even after doing this I don't hold out much hope of truly accurate readings.
These weather stations, in their various brands, are obviously becoming popular. Soon after putting this page on my web site I received an email from Grant Gibson, who'd found this page whilst looking for info on the version he'd just bought. You can read more about his installation on his blog.
Inspired by Grant's efforts, I went to B&Q and bought a 6' long 19mm tube to make a taller mast, in the hope of getting more realistic wind measurements. The tube is meant for use as a wardrobe rail — it's a bit thin walled for this application, and I suspect it may not survive a high wind. (But at least I'll have measured the wind speed at which it collapses!)
Fitting the alternative mast was quite simple. I sawed a single slit, about an inch long, in the end of the tube. Squeezing this slit closed allowed the end of the tube to be inserted into the upper part of the original two-piece mast. I jammed it in tight and then clamped the bottom of the new mast in the TV aerial bracket. I also took the opportunity to tidy up the cables, using cable ties to attach the cable to the mast at 18" intervals. Within a few months of installation the new mast is beginning to discolour from rust, and I suspect it might not last the winter.
During the summer of 2008 the rain sensor stopped detecting any rain. In November I dismantled the rain sensor and found that a spider had moved in and tied up the mechanism with its "silk". A quick cleanup has restored normal operation.
In August 2011 the atmospheric pressure sensor on my first station failed so I bought a replacement WH1081. In November 2013 the replacement's transmitter failed so I bought yet another one — a Maplin N96GY. This is unusual in using dark coloured plastic for all parts, including the sun shield!
In May 2014 I finally replaced the thin, rusty mast described above. I mounted the N96GY wind sensors on a 32mm diameter aluminium alloy tube sold for mounting TV aerials. I used a 32mm to 20mm reducing coupler for MDPE water pipe to join the new mast to the weather station's stainless steel mast, giving quite a neat result as shown in the third picture.
Having a large collection of spare parts allowed me to make some improvements to the wind sensors. I've mounted the anemometer on an extension mast, to reduce any effect its turbulence has on the wind vane. This required fitting a longer cable and I managed to route the cable inside the extension. I also rewired the wind vane and routed its cable inside the mast.
The wind vanes on these stations are notorious for swinging wildly and never giving a steady reading. I attempted to reduce this effect by adding a couple of foam wedges to the vane, giving some "dihedral angle" as used to stabilise model gliders. I added a balance weight to the front to compensate for the weight of the wedges.
I've found several useful web sites, in particular one by Michael Pendec of Denmark. This includes some information on reading the station under Linux. Starting from Michael's C program I've written some Python programs, using the Python USB library PyUSB, to read data from the device. I can also translate data to and from the EasyWeather.dat data file format. The software also processes the data and uploads it to this web site. It creates graphs and tables, such as last 24 hours weather, and posts hourly Twitter messages. This is all done with a combination of Python and gnuplot. I've made the software available as open source via the pywws project.
Initially I connected the weather station to the Asus WL-500gP V1 router (running "Oleg's" firmware) I use to interface my PVR to my home network. However, I found I could not read any data because of a usb_interrupt_read failure. Compiling Michael's software for the router gave the same problem. I then tried a WL-500gP V2 and found that it works. The V2 version of the Asus router has different USB hardware, but I don't know if that's relevant or not. Finally I've installed OpenWRT 8.09 (which uses a 2.6 Linux kernel instead of 2.4) on the V1 router instead of Oleg's firmware, and it all works OK.
To compare my readings with some proper weather stations, here are some links to some nearby weather stations. My location is (N51.36 W0.25).
This is a list of weather station software I'm aware of that is compatible with the Elecsa / Watson / WH1080 stations. Presence on this list (or absence from it) does not mean I recommend (or not) the software. Only you know what your requirements are, so it's up to you to decide what's best for you.