Since posting this last week, I've had a few queries from folks interested in setting up their own station, so I thought I'd share the process that I went through, in the event that there are others who may also be interested.
I'll preface it with saying that prior to doing it I did have quite a number of lengthy email exchanges with Laurie about potentially posting the data as a Seabreeze location. Laurie's experience and insight was both very helpful and full of great practical experience that helped me enormously.
In the end, Laurie and I agreed that for a number of reasons, this site wouldn't be an appropriate Seabreeze feed, primarily because the equipment used is considerably different from what Laurie uses for the Seabreeze data collection/feeds and the hardware isn't nearly as robust nor reliable as the Seabreeze or BOM gear where Seabreeze gets its data. Laurie expressed concern with how often he'd be fielding queries if the station went offline, along with maintenance and support issues, etc. After this experience, I now clearly understand Laurie's concerns and understand his decision.
Having said all that and armed with this wealth of experience from Laurie I decided to have a go at it on my own with one of the primary objectives being to keep the cost as low as possible.
Laurie's Seabreeze weather stations are all professional units meant to survive extremes of weather without having to be replaced regularly. As a result, these units cost upwards of $1,000 once you'd put everything together to make it work.
The setup I've deployed has cost me out of pocket around $175, inclusive. But that's not the end of the story. More later.
If the outdoor component of this setup (sensors) survives two years in the brutal West Australian sun and the relentless Geraldton winds, I'll be happy.
What makes all this possible is a nifty Python/Linux routine written by a great bloke called Jim Easterbrook in the UK and the recent availability of the Raspberry Pi computer. But I'm getting ahead.
Previously, you needed a weather station, a dedicated computer (like a laptop), wires running from sensors to the weather station computer, Cumulus weather software, and all manner of other things that would occupy a bit of floorspace or deskspace in the area around the weather station.
Jim's routines have changed all that and now this little setup only requires a place to sit your weather station head-end display (about the size of your router) and the Raspberry Pi (the size of a deck of cards) in a location anywhere near your internet router. You'll then need an Internet connection to port the data out for display.
So, here's the parts list:
RaspberryPi: --
http://au.element14.com/jsp/bespoke/bespoke7.jsp?bespokepage=e14/common/promotions/rpi-accessories.jsp&ICID=ap-rpi-bplus-hwb-an Weather Station: --
www.ebay.com.au/itm/Pro-Wireless-Weather-Station-w-Wind-Speed-Rain-Gauge-/280751578419?pt=AU_Gadgets&hash=item415e191d33 Weather Station software: --
www.weather.dragontail.co.uk/index.php?page=station_setup Misc:
Rasp Pi housing
USB-USB cable – Weather station to RaspPi
LAN cable – Pi to Router
USB Power point with USB –USB micro cable
8GB SD memory card
Note that the weather station doesn't have to be this exact one linked above -- Jim's software can work with quite a few different ones as you'll learn in his Dragontail page. The one above is the one that I chose. There are many options. This one goes by many names and they're all pretty much the same. Maxkon 1080, 1081, Fine Offset 1080, Maplin 1080, etc. You'll find them for sale all over the place for around $120.
Before you go any further, a couple of pearls of wisdom for you that may save you a lot of hair-pulling and grief. I don't want to put anyone off who'd like to have a shot at doing this but you'll need some basic skills in certain things to be successful.
Firstly, you don't need to be a coder/programmer/computer scientist to do this, but you will need some basic skills in Linux, understanding of basic Linux commands and switches, some basic to intermediate TCP/IP skills and the ability to not be intimidated with working the Linux command line in a terminal window. FWIW, even though I know NOTHING about Python, I've managed to get my station working without having to do any appreciable Python programming or debugging. As they say "YMMV".
Because the Raspberry Pi is a fully-functioning computer you need to access it to load code, OS, etc. But it's really clunky having to plug a monitor, keyboard, mouse, LAN cable etc into this little box. So the solution is to use all this stuff initially to configure the initial environment, then, you set up an SSH access to the Pi by assigning a fixed IP address to it on your network. After this, all you need is a LAN cable to the router and some free software like PuTTy and you can unplug everything except for the power and LAN cables and do all the other configuration and troubleshooting from any machine on your network.
This project ended up taking a LOT more of my time than I anticipated. Even though I managed to get the sensors and the weather station set up and working straight away with no dramas, interfacing the weather station to the Pi and getting the Pi to send the data out to the web pages was a LOT of work, requiring probably a hundred hours of my time before I got to the stage of not having to be rebooting the Pi or resetting something several times a day to get it back on line.
It's now been running with 100% up-time for more than a week without blinking. Fingers crossed.
Jim is very helpful and hosts a Google Groups page where users can post questions but I did my best not to hassle him any more than I had to.
So, with all that in mind, have a go at it. I've learned a lot in the process and don't regret the exercise. But it was a lot of time invested to get it all up and going.
Feel free to PM me if you have any questions, or if they're relevant to the Seabreeze audience, post them as replies to this.
I'll take some photos of the setup and post them later on.
Cheers,
Rob
www.kitesurfing.com.au