It's now 1 year since I first got open3600 running on my GNU/Linux box and got it talking to be WS3600. The first data recorded was at 2005-12-27 14:16 UT, it stuttered a little for a few minutes as I modified a couple of things and then was up and running.
That same afternoon I also started work on the main site.
It's quite a frosty morning this morning. It got down to around -2.6°C at one point. I couldn't resist getting the camera out and taking a few pictures:
Further to yesterday's post about recording my highest pressure yet, the pressure climbed a little more and I ended up with a record high of 1041.9hPa at 23:17.
Still not quite enough to catch Tim's record of 1047.5 hPa back in January this year.
Back in July I added some graphs that showed the last 150 days' worth of data. I wrote the code to generate them rather quickly and have not touched it since.
This evening I've made some changes to how they're generated (got rid of the user of gnuplot's smooth facility and have also added some labels) and I've also added two extra graphs: the wind chill and also the impact of wind chill (the latter being the temperature minus the wind chill).
Today looks like it's shaping up to be another totally windless day. Just like a couple of days ago it's very cold, foggy, damp and gloomy (and, by the looks of things, it's even worse for large parts of the UK today).
I've also had another record for my station today. At 11:09 I recorded a pressure reading of 1040.6 hPa — that's the highest I've recorded since I started keeping data acquired from the WS3600 (the previous record for me was 1037.2 hPa, back in January).
So far today the weather station has recorded no wind. Looking back at my history a day with no wind recorded at all is a very rare event.
The only other time I've had a genuine day with no wind was 2006-06-17 (although yesterday came very close).
Unsurprisingly today and yesterday have been cold (but not as cold as it has been in Scotland), damp and foggy.
I've had a couple of other days with no wind recorded, but in both cases that was due to problems with the system. The first was on 2005-12-29 when snow clogged the anemometer, the second was 2006-07-18 when heat appeared to cause the impeller to stick.
I've just realised that I've now had the weather station up and running for just over a year. I first installed it on the 17th of December last year (and wrote about the installation in some more detail a couple of days later).
I don't quite have a year's worth of data collected yet, I didn't start collecting and keeping data until the 27th of December (that was the same day that I started work on the main site).
Just over a month ago I posted a number of photographs of a really nice cloud display I found when I wandered over to the office. I've now gone back through my CloudCam archives and made a movie of the images taken around the same time. This is the result:
I've just been looking back over some of the images captured by my CloudCam and stumbled on this little sequence:
The stats page on my weather station site has always had an all-time wind distribution table. The idea is to get some sort of idea of where the wind generally blows from at my location. While it does give that information it doesn't really have that "at a glance" feel to it that you really want.
This morning I decided that this data might also work well as a "compass graph":
I've been staying in London most of the week (in Barnet to be exact). For most of this morning the weather was pretty fine down there, a little breezy but other than that we had nice clear skies, very little cloud to speak of.
As I was getting ready to leave this morning, at just after 11:00 UT, a large dark cloud moved in, it started to rain and the wind started to pick up. Suddenly there was lightning, then the rain really started to hammer down, followed by sleet and then hail. Suddenly the wind got very intense, worse than any wind I've experienced before — it was so windy it set off pretty much every car alarm in the street. And then, almost as quickly as it had come, it eased off.
I was amazed at what I saw but didn't think too much of it after that (other than being amazed and impressed, what with me being interested in weather and all).
The next couple of hours were spent on the underground and then the train back up to Lincolnshire and in all that time I never got to see any news. Then, this evening, I turned on the TV to catch up with the news and caught this story.
From what I can tell, looking at the location of the tornado and looking at the direction the weather was moving in, it seems that what I experienced was the remnants of that tornado.
Although I've not done much in the way of weather photography I do try and snap interesting things when I see them. Because of this, and because I want to try and do more weather photography, I've added a weather photography section to my weather station website.
I've added a tools section to my weather station website. The idea is to make available some of the tools I've written that make the station work (at least those that have a chance of being useful to other people — a lot of the code I've written is very specific to my setup and it wouldn't make sense to publish it).
Yet more messing about with the wind compass graph. After having it up and running for a few days I started wondering what it might look like if I created a video of the graph building up over time. This is the result:
After adding a wind compass graph to my site's graphs, and then after improving them (with a bit of help from Tim), Tim's gone and nicked the idea for his own graphs.
Not only that, he's gone and added lots of cardioid graphs too. I'm not sure what they tell me, but they look kind of fun.
After adding the wind compass graphs earlier today I started thinking again about the problem of averaging the wind direction. Being able to produce a useful rolling average helps with the turbulence problem I get with high wind speeds and the compass plots had a minor bug in them when the wind was moving either side of north.
So, after a little bit of searching around on the net and not coming up with anything that useful (maths really isn't my thing so I find it quite hard to make sense of most materials), I hassled Tim and he came up with a couple of useful pointers.
A little bit of hacking on rollavg later and, I think, it's all working. I can now do rolling averages on wind directions and it doesn't seem to all fall apart when the wind is coming from somewhere up north (yes Tim, the pun was too good to miss ;)).
I've now tweaked the compass plots to show both 20min and 60min rolling average values and I've also reintroduced the rolling average lines back into the old wind direction graphs.
Hopefully, this time around, it'll all work sweet.
This morning I added a new type of graph to the graphs for today and yesterday. It's somewhat similar to the wind direction graph but, instead of showing the direction left to right, it plots it as a compass.
Edit: I've now added a compass graph for the last 7 days too.
I also had a bit of a record this morning. At 05:02 I recorded the lowest relative pressure value I've seen since I started running the station. The value was 972.4 hPa, just lower that the last record of 976.7 hPa on 2006-02-17 at 03:37.
As I headed out to work this morning I was greeted with a great display of clouds and had to dash back indoors and grab the camera:
At 10:41 BST today the base station of the WS3600 suffered another self-reset. This only lasted for about 3 minutes in total (including the time taken to reacquire data from the transmitter). The crash log has been updated.
File Under: WS3600, Weather Station, WS3600 Crash, WS3600 Reset.
At 20:27 BST today the base-station of the WS3600 suffered another self-reset. This only lasted for 1 minute so it won't show up in the All Missing Data section of the crash log but I have added it to the Witnessed Events section.
File Under: WS3600, Weather Station, WS3600 Crash, WS3600 Reset.
The WS3600 base station crashed again today. Not like the reset I had the other day, this was a full crash and required a manual reset.
I've updated the crash log.
File Under: WS3600, Weather Station, Crash, WS3600 Crash.
At 12:07 BST today the base-station of the WS3600 suffered another self-reset. Prompted by this I've now created a page on the main site that keeps a record of all known resets and crashes.
File Under: WS3600, Weather Station, WS3600 Crash.
Further to yesterday's wind speed problem: at around 12:50 today the anemometer appears to have started working again.
Let's hope it was just a temporary glitch.
File Under: Weather Station, WS3600, Wind Speed.
Today, as of the time of writing, the WS3600 hasn't recorded a single instance of wind speed above 0mph (the last record was at 2006-07-17 20:17:06 BST with a wind speed of 0.2mph). This isn't so unusual, I've seen periods before when the wind speed has sat at 0 for a large part of the day before it finally picks up. To the best of my knowledge I've only once had a whole day of not recording any wind — and that was because of snow.
I knew the sky camera would come in handy. This morning, it managed to catch evidence of Rods flying around my own back garden! ;)
I've just added a new page of graphs to the weather site: the last 150 days. At the moment it only contains graphs for temperature, humidity, pressure and wind speed but I'll probably get round to expanding that at some point.
The graphs are updated daily, around local midnight.
File Under: WS3600, Weather Station, Graphs, Gnuplot.
So far the revived CloudCam seems to be working pretty well. I'm still doing the odd tweak here and there and sometimes the exposure isn't quite what it could be but it's a lot better than the last attempt at getting something like this working.
This morning a storm came over my location so, as a test, I've made this little video from the captured images:
Back in March of this year I had a play with setting up a CloudCam. In the end it didn't work out too well so I gave up on that setup.
Last night, after trying to revive the rain bucket, I decided to have another go at getting a CloudCam up and running. This time around I used an old Windows XP box and quickly wrote some code in Delphi to do the capture of images. I'm not considering it live yet (hence the fact that it's not on the menu of weather site yet) but CloudCam is up and running again – for the moment anyway.
Ever since the rain bucket died back in April I've been meaning to open it up and have a look inside to see if there were any obvious problems. As often happens, it was one of those little jobs that just kept being put off.
This evening I finally decided to pop the top off and have a look inside.
While searching for something totally unrelated today I stumbled on an interesting little image: a map showing average UK wind speeds. The map is part of the BWEA's wind speed database page.
File Under: UK, United Kingdom, Wind Speeds, Map.
Given that it directly relates to weather it seems sensible to post a pointer to this log entry on my astronomy site.
File Under: Atmospheric Phenomena.
I'm starting to think that the rain bucket had died. The bucket is in a less than ideal location, it's next to a building which will form a pretty good rain shadow so, often, a small shower or some light drizzle tends not to register. Because of this I'm used to getting no-rain readings when I might have expected something.
However, in the past 24 hours we've had a fair bit of rain and in the last hour we had a hell of a downpour (flooding in the road, drains not keeping up, that sort of thing) and nothing has shown on the WS3600.
On top of that, to try and be sure, I've just been outside and tipped the bucket by hand — I could feel and hear the rocker move but nothing has shown up on the base station. This would suggest that either something's wrong with the bucket, the cable or the main transmitter.
File Under: WS3600, Rain Bucket.
It seems my recent optimism was misplaced. At around 09:22 this morning the base station of the WS3600 crashed again. The usual thing happened, there was a beep, the backlight came on and every single part of the LCD display turned on. It didn't reset itself and I had to remove the power to cause a reset.
So, it seems that the solution isn't turning off DCF77 support and the recent long run of working without a problem has simply been good luck.
File Under: WS3600, DCF77.
Last Saturday was an interesting day from a weather watching point of view as I was witness to two different types of vortex.
Around noon that day, while walking around the village, I saw an unusual arc of cloud coming out of a bigger cloud. After watching it for a moment it seemed clear that the arc was, in fact, a rotating tube of cloud. A couple of other people who were with me watched and confirmed what I was seeing. We watched it for about five minutes as it "detached" itself from its parent cloud and head off on its own, finally appearing to "evaporate".
Later that day, probably some time around 18:40 BST, while driving up the A15 to Lincoln we drove into what appeared to be some sort of storm front. There was a very obvious line of cloud with rain falling from the leading edge and curving under the cloud. That was an impressive sight as it was. Then, as we got further under the cloud my wife pointed out an area of tight rotation within it. I had a look and could see it and commented that, if we were lucky, we might see a funnel cloud form. A friend in the car with us then pointed out that, under the area of rotation, dust was being picked up and spun around in the field below. We watched this for about 20 seconds as we drove past and then the movement in the field seemed to cease.
Sadly, in both cases, I didn't have any kind of camera with me and, in the latter case, I didn't think to note the exact time or location of the event. As best as I can recall I think it probably happened around here.
Annoyingly, at the moment, I can't find any online charts for the UK that show fronts and the like for around that time. My suspicion is that what we saw was a cold-core funnel cloud starting to form so it would be nice to check if there was a cold front in the area at the time.
File Under: Funnel Cloud, Vortex, Storm.
Tomorrow it'll be a month since I last had the base station crash. That's quite a significant period of time given that, before now, it was happening every couple of weeks or so.
So, what's changed?
Well, around the time of the last crash I was also having another problem with the base station: every so often the date and time would go way out — sometimes years into the future, other times years into the past. This seemed very odd given that the WS3600 uses DCF77 to keep the date and time correct. Suspecting that the problem was with the DCF77 update I decided to turn it off for a while and see what happens.
Since turning it off I've not had the date and time go wrong and I've not noticed a reset of the base station and I've not had a crash. This could be coincidence (and knowing my luck the bloody thing'll crash the moment I post this) but, so far, it's starting to look like these two issues might be related.
File Under: WS3600, DCF77.
Last Sunday, here in the UK, we turned our clocks forward an hour from GMT to BST. This was a bit of a significant event for my weather station setup because, when I first started work on the tools that would pull the data off it and present it on the website, I decided that local time (rather than Universal Time) would make more sense. For reasons I won't bore anyone with it was important that I selected one or the other as the primary method of dealing with time and given that weather data lends itself more to how we portion up our days local time was the choice that won.
There were a handful of minor problems, all of them were to do with presentation and, in each case, they were to do with the fact that I'd written some code that assumed UT rather than GMT/BST. As far as I can tell each of these issues is now dealt with and things are displaying correctly.
Of course, the real test will be the switch from BST to GMT. Skipping forward an hour is one thing, skipping backward an hour is quite different.
The bits I ordered a couple of days back have turned up. I just quickly stuck one of the propellers on the motor, wired it up to the solar panel and went outside while there was some sunlight about and.... it worked!
That's not to say it's going to solve my problem, but I could feel a breeze off the propeller so in the confined space of the shelter it might just make a difference.
I don't know when I'll get to actually work on this, and I need to figure out how best to hook it all up and how to site it in the shelter — and being the sort of person who knows very little about electrics and related matters it might take some time.
File Under: WS3600, Shelter, Temperature, Thermometer.
After this morning's apparent failure of the shelter to solve the temperature reading problem I got to thinking some more about how the problem can be solved. One option would be to have some sort of fan that would keep the air moving in the shelter — I don't know how well this would solve the problem but it's something that's got to be worth a try.
So, how to go about it? I guess the obvious solution would be a PC fan with a power supply and I'll probably look into that at some point. However, I've been looking for an excuse to order some fun things from MUTR so I had a quick look for some ideas on there. One little shopping trip later and I'm wondering how much fun can be had with a solar panel, a small motor, a polythene fan and a miniature impeller blade. I'll find out soon enough.
Heck, there's even a chance that mucking about with such toys will result in some sort of solution.
File Under: MUTR, WS3600, Solar Power.
Almost as soon as I comment that I've not had a test of the shelter yet the Sun comes out. It's quite cloudy out there and, even in the gaps, it's thin cloud, but there is sunlight falling on the shelter.
And, so far, it looks more of a failure than a success. As of the time of writing (around 09:13 local time) the temperature has gone up to 6.1°C from a starting point of around 4.5°C (and it appears to still be rising).
Looks like I'm back to the drawing board again. I guess the next stop is to look into the idea of turning the shelter into a fan aspirated shelter.
Oh well, at least the transmitter is protected from the elements...
File Under: WS3600, Shelter, Temperature, Thermometer.
Yesterday I finally got around to installing the shelter and moving the thermometer (which is also the hygrometer and transmitter) into it.
It wasn't such a difficult job either — took just under an hour. To be on the safe side I disabled the downloading of data from the base station while I did the work (you can see the gap in the data here). I did this because I was going to be handling the thermometer and also unplugging the anemometer and the rain bucket — given that no data would be coming from them there was a chance that odd values could have been recorded. The only error to come out of the exercise (so far anyway) was a record of 0.5mm of rain; chances are I nudged the rain-bucket while getting the ladder in place and this would have caused a tilt.
So far, so good. One concern was that putting the thermometer in the shelter would mean that it didn't record lower temperatures so well but the values recorded over night look pretty much okay. Right now the external temperature is showing as 0.9°C — exactly the same as the little Maplin internal/external thermometer is showing. My nearest personal weather station is currently showing 0.4°C.
The real test will be what happens when it's sunny and sunshine is falling on the shelter itself. It's totally overcast at the moment so it looks like I won't be able to test that today. Unsurprisingly the forecast for a large part of this week seems to be more of the same so I might have to wait a while to see if the shelter solves the problem.
File Under: Personal Weather Station, WS3600, Shelter.
After this morning's problems the camera seemed to settle down a little bit. The image has been far too dark all day but most of the time you could see the clouds and the light levels of the images did seem to relate to the light levels outside.
So, now that I've got a whole load of jpeg files sat in a directory, what to do with them? Easy, take one copy of ffmpeg and make them into a movie!
Note that it's a .mov file and that it weighs in at around 2.2MB.
If I can get the camera to actually behave itself and if I get an interesting weather event happening this might be one interesting additional way of recording it.
File Under: Weather, Clouds, Sky Camera, Cloud Camera, WebCam, ffmpeg.
The first day of testing the camera more or less worked and, unsurprisingly, over night all that was shown was a black square. However, this morning, it's not worked very well at all.
Yesterday I initially started using a tool called streamer to grab the images from the camera. When I had it set up in the office, pointing in the office, it worked fine. However, when I had it looking out of the window I mostly got a blank, white image. I then added a delay (this can be supplied on the command line) to give the camera time to adjust its brightness levels and that more or less worked. But, even then, I'd get the occasional white-out image.
I then decided to have a look around and see if I could find a different command line based tool for grabbing an image. vgrabbj was my next choice. This appeared to work rather well and it produced good images without the need for a delay. It also seemed to be a better choice because it provides a method of adding a timestamp to the image itself.
This morning, from just after sunrise, it's also been producing mostly white-out images too. It also has a parameter that causes a delay in grabbing the image and so I've made use of that. Even then the camera still seems to be having problems adjusting to the light levels. Chances are the problem all comes down to it being an old camera.
I'll carry on playing with it for a short while longer but, at the moment, the current setup isn't looking too promising.
File Under: Weather, Clouds, Sky Camera, Cloud Camera, WebCam.
Today I dug out an old Creative VideoBlaster II web camera; the old sort that connects to the parallel port and takes power from the keyboard connection. I hooked it up to one of my GNU/Linux boxes and rebuilt the kernel to provide support for the camera.
So, for the moment anyway, my weather site now has "CloudCam".
It's simply sat on a windowsill, pointing out of the window at the sky. To be honest it's not a very good setup and the quality of the image is pretty terrible (to take an image I have to start it up and then wait 12 seconds before the balance has sorted itself out) but it's a fun thing to try out. In the long term, if it does work out okay, I might see about hooking up a better camera and sorting out a better position.
File Under: Weather, Clouds, Sky Camera, Cloud Camera, WebCam.
Last Saturday I happened to pop into the office and, while in here, noticed that the base station had crashed again. This happened at around 14:32 and I revived things around 14:57. It would appear that the crashing of the WS3600 is going to be a pretty regular event.
So far, since the station went online on 2005-12-27 I've had the following problems:
Between 05:00 and 05:30 this morning the weather station recorded the lowest temperature I've managed to record so far: -6.6°C.
I was actually out doing some observing at the time and it really did feel rather cold.
Sadly, despite the fact that as of the time of writing it's still around freezing out there, the station is recording 12.1°C due to sunlight warming up the thermometer. I'll be really glad to get the shelter finally finished and installed. Just one part of it needs a final coat of paint and then I can get it all set up.
File Under: WS3600, Record Temperature.
Yesterday evening I finally got round to creating a database for my weather data (using MySQL) and writing some code that injects the data into the database. While the site itself has never been, and probably won't ever be, driven off a database doing this does help me improve the content of the site.
Having the data in a database that can be accessed with SQL means I can do all sorts of queries that'd be very tricky when working with the log files I create with Open3600's logging utility.
The first visible outcome of this is I've started a stats page on my site. There's only a couple of tables on there at the moment — I'll probably add more as other pointless ways of looking at the data come to me.
Next on the list will be re-writing the records page so that the data it uses is created from the database and so that more time periods are included. Tables for current week, current month and current year will probably happen at some point soon.
File Under: WS3600, Open3600, MySQL, SQL, Weather Database, Weather Statistics.
About an hour ago it got very dark outside and the wind picked up, followed by a very short and sharp hail storm.
Just like a few days ago, the WS3600 base station had another crash yesterday &mdash total lock-up, every part of the display turned on, nothing working.
Sadly, this time, it didn't happen while I was at my desk; it happened on a Saturday. The only reason I noticed is because I happened to check the current conditions using my PDA and I saw that an update hadn't happened for a while. Because of this no readings were recorded between around 14:00 and 16:40 (you'll see the gap here).
This business of resetting or, worse, crashing, is starting to annoy me a little now. It suggests that the WS3600 can't be relied on to work unattended — it looks like it needs lots of care and attention. I've done a fair bit of searching on the net for any information on this problem and can't come up with anything useful. At this rate I might see about getting in contact with La Crosse Technology to find out if they're aware of this problem.
File Under: Personal Weather Station, WS3600 Crash, WS3600 Reset.
Another one for the "because I could" department: I've been having a bit more of a play with gnuplot and have ended up creating 2d and 3d plots of wind direction vs wind speed.
For the moment they only appear on the page of graphs for yesterday's data; the reason for this is that creating these plots is a little more processor-intensive than the other graphs and, if I create them for the last 7 days they look rubbish.
File Under: Weather Data, Wind Speed, Wind Direction, gnuplot.
Just over a week ago I added a simple XML feed of current readings to my weather website. At the time I didn't really have much of a need for it, I just did it as one of those "because I can" things.
This morning I got round to writing my first client application that makes use of that data. davep-wx.el is a little utility for use in GNU emacs; it provides a command that displays the current weather readings. Simple, straight-forward, but handy. Also fun to hack together (I do like writing code for emacs).
File Under: GNU emacs, emacs lisp, elisp, Weather, WS3600.
Like last month, I've just had the base station of the WS3600 reset itself for no obvious reason. As with the last time this happened, there was a beep, the backlight came on, every part of the LCD display turned on but, unlike last time, it didn't actually reset — it just seemed to hang.
After a short wait I finally pulled all power, waited a moment, and then hooked up the power again. This time the base station started up okay and, a short while later, data started to be received from the external sensors again.
This time the reset appeared to retain the rain total, the corrected pressure offset and the date and the time (although the time was the time that it had hung, not the current time). Everything else (the min/max values) were all lost.
If this keeps up it might start to get a little annoying and frustrating. Worse still, I know of one other person who has all sorts of odd problems with his WS2300 and I've had an email from one person, with a WS3600, who was having these sorts of resets on an almost daily basis. Sadly, I don't seem to be able to find any sort of explanation for them.
File Under: WS3600, Personal Weather Station, Strange Reset.
Further to last week's update, I managed to get a second coat of paint on the shelter over the weekend. Thankfully this coat went on a lot better and it's starting to look the part now. There's still more painting to be done, I've still got to do the inside of the door (it's pretty much impossible to paint the inside and the outside at the same time without getting paint and fingerprints everywhere).
File Under: Personal Weather Station, WS3600, Thermometer Shelter.
I've had a long-running (for "since I've been using this setup" values of "long) problem with the WS3600 and Open3600 whereby, now and again, a pressure value of 11111 hPa would be recorded (first noted back in December last year). At first I wondered if it was a problem with the base station itself but, because this high value was never being recorded as a max pressure, I concluded that it was probably a problem with Open3600.
Since then I've been getting, on average, one instance of this problem every couple of days.
The base station is say on my desk and the display is visible out of the corner of my eye as I work on my computer. Earlier today, as I was working, something caught my attention and I looked over at the base station and saw that the pressure was flashing "-----" rather than displaying a value — the rest of the display appeared to be fine. A moment later a proper value was displayed.
I then checked in the data that had been recorded and saw that another instance of 11111 hPa had appeared. It would seem that it isn't Open3600 that's at fault but it is, instead, an issue with the base station.
Knowing that doesn't actually solve anything, but at least I now know what the source of the silly value is.
File Under: WS3600, Open3600, Bad Pressure Reading.
I've noticed that some personal weather station sites, especially those that use Weather Display, display the height of the cloud base. I've often wondered how they got that figure.
Assuming that there was some sort of calculation that could be done I did a search on Google and came up with this and this. Both sites give pretty much the same method of calculation.
So, more for fun than anything else, I've added the value you get to various parts of my website. The tricky part is finding out a way of checking if it's anything close to correct (for Cumulus clouds, obviously). I doubt I'll be able to do that any time soon (doubtless there are some neat gadgets you can buy that'll let you do it).
Of course, I made the silly mistake of telling a certain Scottish-based geek about this and he just had to add it to his weather station site too (actually getting it in before me). But that's okay, I'm letting him off this time because he was kind enough to let me illustrate this blog entry with the image you see above.
File Under: Personal Weather Station, WS3600, Calculate Cloud Base, Cumulus.
Just an update on how things are going with building the shelter. Over the weekend I had my first chance to get a coat of paint on it. I did this on Saturday afternoon and, by Sunday evening, it still wasn't quite dry. It's also going to need at least two coats (perhaps even three) — which is kind of odd considering that I'd grabbed a "single coat" paint to do the job.
I'm hoping to get the extra couple of coats on it this week so, fingers crossed, I might have it all done by next weekend. If I have the time and the weather is good I might be able to get it fixed up outside. The chances are that I won't get the time to move the thermometer itself and I'm tied up the weekend after next so it might be over a couple of weeks before I've got everything in place.
Thankfully, for the last few days, it's been overcast all the time so temperature readings have been correct. Of course, wearing my astronomer's hat, I'm sick of this overcast weather.
File Under: Personal Weather Station, WS3600, Thermometer Shelter.
Tim Haynes, a friend and fellow geek that I've known for some years now, acquired a personal weather station not long after I got mine and he's also been setting his up and building a website based around the weather station.
This last weekend he knocked up something very silly but also lots of fun: a Weather Multimedia section for his site. Simply put, he's taking some of the data from his weather station and producing sound files from it.
File Under: Personal Weather Station, WS2300, Podgraphing, Sound Files.
I've now started work on the shelter. So far, this afternoon, I've managed to build the main box and put most of the door together. It's turned out to be a little harder than I thought it might be. I'm the first to admit that I'm more than a little lacking in the DIY department (and I'm also one of those people who gets a little frustrated when things don't go together as easily or as well as I'd hoped) but this seems a little tricky — trickier than installing the weather station itself.
It's not that there's obviously anything wrong with the kit (not that I'm really in a position to know if there is), it's more a case of it being a job that often needs two pairs of hands and most of today I've only had the one pair.
The more it comes together the more concerned I am that it might not really do the job. The problem I've got is that the only wall I can mount things on faces east (ideally I should be mounting on a north-facing wall) and, obviously, the Sun in the morning causes me problems. I need something that's going to fully shade the thermometer during the morning.
Having seen the size of the gaps there are in the door (see right) I'm concerned that, during the morning, there might be times when the Sun is managing to shine right into the shelter and cause the problems I'm getting now.
I'm going to carry on with the build and installation no matter what (the install will probably happen in a week or so — once I've got the shelter built I still need to paint it and I generally only have the free time for this sort of thing at weekends). Even if this turns out not to be the perfect solution it should, at the very least, help protect the thermometer from the wind and rain. Also, if it doesn't fully work, the chances are that it gives me a base to build on.
File Under: Personal Weather Station, WS3600, Shelter, Thermometer Screen.
This morning I added two new facilities to me weather site. The first addition is a simple RSS feed of current conditions (if you've got a feed reader you'll find the feed over here). A sample of the current conditions is taken every ½ hour and added to the feed.
The second addition is a simple Google personalised homepage module. This gives a simple text display of the current conditions (the display is taken from here). If anyone wants to have a play you'll find the module here.
The latter seems to be all the rage today. It's nice to know that, where I lead, others follow. ;)
File Under: Personal Weather Station, WS3600, Google Homepage, RSS Feed.
For the first time since I added a "rolling average" plot to the wind direction graph we've had a wind from the north. This, as you'll see from the image to the right, has shown up a rather obvious problem with plotting a rolling average of wind direction. Because the wind direction oscillates around 0° I end up with an average plot that suggests that the wind is coming more from the south.
D'oh! (Yes, imagine me slapping my forehead in a "it's really obvious now I see it" way)
Mostly the average plot works really well and nicely shows where the wind is generally coming from so I'm left wondering how to correct this.
Now I can see why some of the more popular weather station software plots wind direction on a radar-a-like graph.
I've been meaning to see if I can do anything with the polar plotting facility in gnuplot — this would seem to be a good enough reason as any to have a play with it and see if I can produce something sensible.
File Under: WS3600, Plots, Charts, Graphs, Wind Direction.
Starting this weekend, I've ticked off another item on my weather station TODO list: I've started to submit data to Weather Underground. This means that my station has a page over there and you can also see it on the map of stations in my area.
Because I'm using Open3600 to pull data from the station I didn't have a tool to hand to do the submission (there is a tool that's part of the package — apparently inherited from Open2300 but it doesn't compile and seems to need a fair bit of work) so I decided to write my own submission program. I went looking for documentation but couldn't find any. However, there seems to be enough scripts around on the net to do this sort of thing that it was easy to see how WU accept the data — you simply supply it via an easy-to-construct URL. Half an hour of hacking away in ruby and I had a working tool.
I've set things up so that data is submitted once every five minutes — I figure that this strikes a nice balance between producing too much data and not producing enough.
The only outstanding jobs now are to see about constructing the shelter for the thermometer (which, I hope, will cure the temperature spikes I get when sunlight falls on it) and to find a better location for the rain bucket (it's currently in a bit of a rain shadow).
File Under: Weather Underground, Open3600, Open2300, ruby, Personal Weather Station.
This morning I finally got around to sorting out what to do about the pressure setting on the base station. I did some more searching on Google and found this thread on this BB.
Following the advice in there I adjusted the relative pressure setting in the base station's configuration menu, setting it to the current absolute pressure reading plus 1.25hPa and this seems to have put me pretty much in sync with other stations near me. It does, of course, mean that there's another large jump in the pressure data (you can see this from after 10:00 in this table) and that all previous data is "wrong". I might have a think about going back and fixing it at some point. That said, I did imagine that the first month or two of running the station would be a "play and see what happens" time so I'm not too worried about it. Given that I've still got to shade the thermometer and I need to site the rain gauge where it's not in a rain shadow I'm happy to take the view that this setup isn't fully "live" yet.
File Under: WS3600, Relative Pressure, Absolute Pressure.
The bits that I found were missing from the shelter arrived today so it looks like the only thing that's stopping me from getting it built and set up is free time.
File Under: Metcheck, Instrument Screen, Personal Weather Station.
Further to the initial crack at getting a history section working, I did some more work on the site this morning and polished off the history section. Now data for any day can be viewed. The default is to view a hourly snapshot but other intervals, including viewing per-minute, are available.
Next job on the site is to include the rain data in the graphs and in the history.
File Under: Personal Weather Station, WS3600, Weather Website.
The shelter I ordered last week arrived this morning. I've not had the chance to check all of the contents yet but, at first glance, it looks like it's more or less all there.
I say "more or less" — there's a note attached to the instructions saying that the instructions are from the 1980s and, since then, the wood they use has changed. Because of this it claims that the panel pins that are listed in the inventory are no longer included and that you should simply use a glue (they suggest No More Nails). Something else that's in the inventory but doesn't seem to be in the pack is the latch that holds the door shut. Hardly an expensive item to go out and buy but I'll drop Metcheck a line just to be sure.
Update: I just realised that the latch for the door isn't the only item that is missing — the hinge for the door is also missing. An email has been sent to Metcheck.
Update: Just had an email back from Metcheck. The latch and hinge will be going in the post tomorrow.
File Under: Metcheck, Instrument Screen, Personal Weather Station.
I've been meaning to sort out a screen for the thermometer/hygrometer for a couple of weeks now but, because the weather's been overcast most days for a while and the over-reading problem hasn't cropped up it sort of slipped down the TODO list.
Today's been quite a sunny day and the problem came back:
I'm confused — horribly confused.
I finally decided to get around to sorting out the relative pressure setting on the WS3600. I dug out an OS map of my area and checked my height above sea level (about 10 meters) and then, using this article to guide me, worked out what the relative pressure setting should be (around -1.25hPa based on that document if I'm reading it correctly).
So far so good.
According to the manual that comes with the station it says that I should subtract that value off 1013.0hPa in the "Relative Pressure" setting in the unit's configuration menu. So, I dived into that configuration item only to find that the value in there was something totally different. After a little head scratching I realised that the value I was seeing in there was the current relative pressure reading — not some sort of base value that I needed to fiddle with.
I made an attempt to adjust the value anyway (as you'll see starting from around here) and I think it's changed things in the right direction but it hasn't actually got it right yet.
As of the time of writing my station is showing an absolute pressure of 1019.9hPa and a relative pressure of 1011.2hPa. I've been wondering if it will suffice to simply set the relative pressure to 1019.9 - 1.25 (in other words: current absolute minus offset) — while that probably still wouldn't be spot on I get the impression it would be closer to correct. I've been looking at the values for stations near me (as of the time of writing: RAF Cranwell showing 1023hPa, Grantham showing 1022.3hPa and Branston showing 1022hPa) but I'm not sure if I should be reading those values as relative or absolute values. All are close to the absolute value I've got at the current time (and, of course, I'm close to sea level).
Like I said: I'm horribly confused now.
I think I've got some more reading and learning to be doing...
File Under: WS3600, Relative Pressure, Absolute Pressure.
This morning I added a simple WAP page to my weather site — using this page (unless you've got some sort of WML browser plug-in for your web browser that link isn't going to work too well for you) I can now check in on the current conditions at my station from my mobile phone.
Not that it's that useful (most of my time is spent sat next to the station), it's more a case of "I can do it so I've done it".
File Under: Weather, WAP, WML.
I've added another page of graphs to my site. In this case it's a set of graphs of the previous 7 days of collected data.
File Under: Personal Weather Station, WS3600, Weather Graphs.
I think I've just noticed another effect of strange reset. When I sat down in the office this morning I noticed that, while the time on the WS3600 was correct, the date was showing as 2006-09-06 — a full eight months into the future.
Thankfully this doesn't seem to have affected the data I'm pulling off the base station. From what I can tell (I'll be checking the sources later) Open3600's logging utility writes the time stamp to the file based on the time on the machine running it, not based on the time on the base station. The only visible effect this has had is that some of the min/max values that are shown on the opening page of my site appeared to be in the future. I've reset the offending min/max values.
I had a look in the WS3600's manual this morning and, annoyingly, there doesn't seem to be any sort of "trouble shooting" section so I'm no wiser about the possible cause of the reset. Later on I'll do some searching on the net and see if I can find any more information.
Update: I just noticed that the minimum windchill recorded by the base station was on 2006-01-06 at 00:33. This is a long time after the odd reset. This would suggest that the date went out of sync a long time after the reset. Now I'm really confused.
File Under: WS3600, WS3600 Reset, WS3600 Incorrect Date, Personal Weather Station.
