In April ’99 I presented a plan to the WWPUG and KWARC executive to install and connect an automated weather station on the APRS network. This project would raise our profile by providing real-time weather observations both locally and worldwide over RF and Internet. The funding of the plan was accepted enthusiastically and the project was afoot.
The weather instrument package selected was a Peet Brothers Ultimeter 2000. That unit was selected because it is highly respected in the APRS community, besides being the only one that can provide packet data to RF without the aid of a computer. The data it provides in one short packet burst includes: wind average/gust, wind direction, outdoor temp/humidity, rainfall, barometer, 3-hour barometer change and date/time – all with remarkable precision. Paul Cassel VE3SY assisted by ordering the unit at the manufacturer’s booth at Dayton. The unit was received in July and within a week was up in a temporary location with assistance from Tedd Doda VE3TJD. At the same time, I replaced the TNC running the APRS node with a Kantronics KPC-3 (these have become the de-facto standard for APRS digipeaters) and re-christened it VE3KSR to reflect its permanent location. It became obvious fairly quickly that the system was severely challenged by the intense RF environment created by the broadcast (and other) transmitters on the hill. The early days of the installation were spent installing toroids to attempt to control the RFI as well as installing a data extension cable to allow the unit to be mounted in its final, clear-air location. After installing toroids in every conceivable location, we still had RFI problems: the humidity reading was very erratic, often going into the low single digits (which would only be reasonable on the surface of Mars). Also, the unit would keep eating up its backup battery – obviously due to RF coming in the DC power. I entered into email conversation with the manufacturer to see if the RFI filters they sell would be effective and where they would best be placed. At the same time I also began work on a program to run on the KW-IGS web server where the club’s web site is hosted. The idea was to capture the system’s data packets that were being beaconed on the APRS frequency and gatewayed to an Internet backbone. The program started out very rudimentary, just parsing and displaying the current observation packets but, as you’ll see later, has grown quite a bit. One thing became obvious fairly quickly – the Internet was a pretty unreliable way of getting the data from Baden to Waterloo. Frequent outages from various causes, usually at the most inopportune time (like when a storm is approaching and in progress) convinced me that we needed to install a local gateway. Although we had an offer to host the gateway at the home of Don Sinclair VE3ICD (the KW-IGS proprietor) I began playing around with Linux, some free APRS server software and a Sympatico High Speed edition full-time internet connection. This is now pretty reliable with various tricks being performed on both ends to keep the data flowing. In the meantime additional RF filters were ordered from the weather station’s manufacturer. All the connections are made using RJ-11 or RJ-45 style connectors and I was initially dubious because the filters were just toroids and capacitors mounted inside phone jack boxes. However, where I had only achieved a few turns through a snap-on toroid when winding the whole cable, these filters carefully attached each conductor to a length of magnet wire that was then much more flexible (and smaller, due to no plastic insulation) which allowed many more turns. The manufacturer claimed 30 db suppression and they did the trick! One at the control unit got rid of the battery eating problem and one at the instrument package fixed the humidity sensor problem. I now focussed my attention back on the web site interface. I quickly added calculation of windchill and humidex (although that probably won’t show up until next year – it doesn’t display unless the temperature is over 18 degrees). I also added a history buffer to keep data for the current and previous day. This allowed me to start capturing daily high and low values. Finally, I added graphs of today and yesterday’s readings. I know that at least one person was watching the page evolve over several days as I learned to use the drawing package – first with one curve, then more, then colour and finally numbered scales. It’s the graphs that I find fascinating. You can actually see the movement of weather systems by watching the lines dance around, for example: watch the temperature suddenly drop and the humidity rise as a shower goes through, see exactly when the rain falls and with what intensity. See for yourself at http://www.kwarc.org/ksrwx.html . What’s next? Oh, I have a little idea that we may be able to make it speak the conditions on demand on the VE3KSR voice repeater – but I need to catch my breath before starting that one!
|
