I had read that the Guruplug wasn’t particularly reliable as a wireless network client.  There are instruction on making it behave as a reliable client, which I had followed when I set up the wireless networking.  However, today the wireless networking hung.  I have two network connection to the Guruplug.  I have the wireless connection, and I also have a wired connection to a router configured as a wireless bridge.  I could still log in using a wired connection, so I could see the Guruplug wasn’t dead.  Unfortunately although there were two network connections to choose from, it insisted on using the hung wireless connection.  So I edited /etc/network/interfaces to disable the wireless on the Guruplug.  I’ve used the bridge for several years and had no problem with it, so hopefully the hardwired connection will hold up.  I’ll let it run and see how it does.

We had a pretty good rain this week.  I was watching the readings from the rain gauge and they were looking very good, increasing by no more than a few hundredths at a time .  All of a sudden the rain count jumped to 8 or 9 inches, then to 14 inches.  It’s very embarrassing because that amount of rain got reported to CWOP and all the other services my data goes to.  My rain readings are probably being blacklisted now.

My assumption is that water got into the electronics and/or wiring.  Either that or the pivot for the tipping bucket mechanism got stuck.  I plan to take the rain gauge apart this weekend, make sure everything is clean and moves freely, and add some additional caulk and weather sealant to make sure the electronics stay dry.  It just figures, as soon as I get the temperature working something else has issues.  At least I won’t get bored or run our of things to do.

I’ve been using my smart phone to check the status of my weather station while I’m out of the house.  This works OK, but required me to zoom in and scroll around on the pages on my weather Web site, as they were too large to fit on the mobile device’s display.  I decided it was time to create a dedicated page for my mobile device.  I put together a quick prototype that seems to work well.  It even has Ajax updates and colors the changed values green, just like my main page.  It doesn’t have much in the way of a flashy user interface yet.  Just a very basic display of several readings from the weather station.  I’ll probably add more to the page as time allows.  The page is at http://www.cloppermillweather.org/mobile.php.

I also built a prototype for a new Web site using a different CSS template.  The layout is very different, and I’m not sure whether I like ti or not yet.  It has a much smoother look to it, but it is more for a blog than for a site that displays tabular data.  I’m going to continue to experiment with it, but I haven’t decided whether I want to make the switch or not.  This new page only uses static data for layout purposes, and can be seen at http://cloppermillweather.org/testsite/

Finally, the revised FARS has been working out perfectly.  Comparisons with local stations and my extra thermometer shows that my temperature readings are always close to the others.  There is still a little solar heating, but it is greatly reduced and not really noticeable unless you know just when it is occurring.  I know because i have a solar detector that can give me an idea as to when the FARS is getting direct sunlight.  The solar detector isn’t much more accurate than that, but does help me see how the temperature is affected by the sunlight shining on the FARS.

I added a new page to the Web site that allows viewing of the last 1440 readings from the weather station in a tabular format.  This number of readings should equate to 24 hours worth of data at 1 row of weather data collected per minute.  That is assuming there was no downtime during that period.  I then added a  link that provides the ability to download that information as a comma separated values file so that it can be loaded into a spreadsheet or database for analysis.  I later learned that PHP has the ability to directly create PDF files.  I may choose to add that feature as well, but I’m not sure how useful it really is.

On a side note: The revamped FARS still seems to be working well.  I’ve been comparing the temperature readings to my other thermometer and to surrounding stations.  The readings dipped down much lower than most of the neighboring stations last night, but the other Germantown station and my other thermometer agreed with the low readings.  I will continue to monitor, as this hasn’t been enough of a test yet.

After running for a few days with the fan off, I’m no longer convinced the fan is the whole problem.  The temperature reading still doesn’t always make sense.  I received the new fan, and I plan to install it tomorrow.  I’m also going to increase the distance between the flower pot sections from 1 inch to 2.  This to allow for more air flow.  I’m thinking in addition to the fan getting hot from the over voltage, that I’m just not getting the air movement through the housing that is required.  The flower pots are a fairly tight fit when they aren’t spaced far enough apart.  Hopefully these two changes will achieve the desired result.  I’ll post an update after I make the changes, and again after I let things settle for a few days.

Update: I installed the new fan this morning and it works fine.  The rating on this fan is 3700 rpm at 24VDC.  The old fan was rated for 12VDC, but the power supply feeding it is rated for 20VDC.  This caused the fan to spin a lot faster than it did at 12VDC, and that was what led me to believe it may have been running hot.  The new fan is actually rated for more voltage than the 20VDC I’m feeding it, so over-voltage shouldn’t be a problem.  The other concern is whether 20VDC is sufficient for the motor to start.  I tested it with a variable power supply on my bench before installing it and it took about 15VDC to start spinning. So 20VDC should be plenty of voltage to run the fan.

I also spaced the sections of the FARS farther apart to help the airflow.  I created some 2 inch spacers to replace the 1 inch ones I had been using.  The extra length required me to remove several sections that weren’t essential for it working correctly.  Then I mounted everything back on the pole outside.

Initial comparisons with my other outside thermometer and nearby stations on Weather Underground look great so far, with little or no solar heating taking place.  I need to let it run for several days to see how things perform over time, as other changes I made in the past appeared to be working at first.  That could have been caused by my opening the FARS up and releasing any warm air that had built up inside of it.  Over time the warm air would build back up again due to poor ventilation.  Maybe I really got it right this time.  I’ll know better after a few days.

For the last few days the analysis of my temperature reading showed it as being high compared to nearby stations.  Sigh, here we go again.  I’d been watching the temperature readings along with neighboring stations on Weather Underground and had come to suspect my readings.  So I started doing some investigating to find whether I had a problem or not.

First I opened up the FARS to have a look inside and make sure everything was OK and still in place.  I didn’t find any problems there.  My humidity sensor has a thermometer in it that isn’t as accurate as my main temperature sensor but could be used for comparison.  It agreed with my temperature sensor, so I was able to rule out a sensor malfunction.  I have another digital thermometer located in another part of my yard.  It was reading several degrees lower than the weather station.  So I knew I had some kind of problem inside the FARS housing.

Since I had some problems in the past with a voltage regulator getting slightly warm on my temperature board, I removed the jumper that supplied power to that regulator.  It is only used to supply external power for the 1-Wire sensors and isn’t strictly necessary since the sensor can run using parasitic power.  I waited and watched the temperature readings, but they didn’t change.  So now I need to put the jumper back on the board.  Oh well, no rush.  It seems to be reading fine using passive power for now.

I thought I might still have a problem with the fan in my FARS generating some heat.  I tried adding a potentiometer outside of the housing to slow the fan down, but that didn’t seem to help.  I had also tried just disconnecting the fan altogether, but I wasn’t certain that was the problem or not.  In testing i had done in the past it had appeared to be working the same regardless of whether the fan was plugged in or not.  After giving it some more thought, I decided to turn off the fan again and observe the behavior.  .If I was right about the fan causing some heating, the expected behavior would be a spike in temperature immediately after shutting off the fan due to the heat generated by the motor not being evacuated from the housing by the fan.  Then as the fan cooled off I should see the temperature fall to match the neighboring weather stations and my digital thermometer.  This wasn’t what I had observed the last time I tried, but this time the fan really looked like the most likely culprit.

I unplugged the fan and left for work.  Later in the day I checked the graphs from my station and saw a spike in temperature just after I unplugged the fan, then a gradual fall in temperature.  Eventually the temperature dropped to where the daytime temperature from my station more or less agreed with the neighboring stations. The nighttime temperature was significantly lower than most of the neighboring stations except for one.  But it also agreed with my spare digital thermometer.

Looks like my fan heating theory was correct.  I’ve left the fan disconnected for now.  This means I will probably get some solar heating in the morning when the sun shines on the housing.  I can live with that for now.  And I plan to order another fan in the next few days for installation next weekend.  Tine to hit the catalogs and research fans.

I registered my weather station with AWEKAS today.  This is another online service that aggregates and provides weather data on the Internet.  They have a very simple interface for publishing data so I went ahead and added it.  They come to a URL on a periodic basis that you provide and pick up a small data file with your weather data.  I built a PHP page that returns the weather data in the expected format with a content-type of text/plain.  So the data file is virtual and built on demand.  So far they are receiving my data OK.  There really is no need for me to publish my weather data to so many different sites.  But it’s all part of my enjoyment of the hobby.  Which hobby you might ask?  The weather station, software development, electronics, or abuse of PVC plumbing and conduit?  Definitely some combination of the above.

I finally got around to adding a ground wire for the lightning sensor.  I had a spool of solid 16 gauge wire in my collection that fit the bill perfectly.  I attached one end to the 4 foot copper ground rod I already had installed.  Then I ran the wire across the ground and up the weather station pole to the lightning sensor.  I attached the wire to the pole using wire ties.  Where the wire ran across the ground, I secured it by cutting short pieces of wire and bending them into staple shapes.  I then pushed these into the ground over the lightning sensor ground wire to secure it.  By Spring time, the wire will be embedded into the lawn.

During the process of adding the ground wire, the lightning sensor detected a few false strikes due to my attaching the wires without shutting off the sensor.  This pointed out a few problems with the lightning stats on my Web page which I quickly fixed.  Now I just need to be patient and wait for some real lightning.  That could be several months, as we don’t typically get a lot of lightning during the Winter, although it has been known to occur.

I looked a little more at new CSS templates for the Web site.  But I’m getting more interested in setting up a WEFAX receiver.  I have an ICOM PCR-100, which is capable of receiving the signals.  However I don’t know that it has sufficient bandwidth to get quality images.  But in the interest of cost and experimentation I plan to connect an appropriate antenna to it and see how it does.  I can choose to add a dedicated WEFAX receiver later if it goes well and I decide this is something I want to pursue further.  I found some plans for home-built antennas made using PVC tubing.  That sounds right up my alley.

I found a copy of the ARRL publication Weather Satellite Handbook by Dr. Ralph Taggart.  It’s out of print and some of the information is pretty dated.  But it does have plans for building some of your own equipment.  I even ran across some special interest groups on the Web for weather satellites that sell receiver and antenna kits.  Although no new WEFAX satellites are being launched, the ones in service should remain that way for a number of years to come.  So as dated as this part of the hobby may be, I can still get quite a bit of enjoyment out of it.

Most of my time on the weather station recently has been monitoring the station and updating the Web site.  I’ve been going through the logs on both my weather server and my ISP looking for problems.  The good news is that I haven’t found any.  I set up logrotate on my server to keep the size of the logs in check.  So far all the logs and log maintenance look good.

I’ve done some minor clean up and additions for the Web site.  I updated the “About” page with the most recent information.  I added some pictures and a few colored boxes to spruce up the page a bit.  Some of the other pages needed some minor revisions, like making the pictures the same size and adding some borders.  On the “Advisories” page I added hazardous weather outlook and the weather discussion from the NWS.  I may decide to take some of that information back out, as I don’t want to clutter things up with extra information that can be obtained directly from the source of record.  Also, they weren’t from XML feeds, so I’m dependent on the NWS not changing their page layout.  Otherwise my PHP code to retrieve that information will need to change every time the NWS changes their page layout.  Not good for maintenance.  But it did give me some practice with regular expressions in PHP.

Tomorrow I hope to get to the ground wire I keep forgetting about.  I’ve also started looking at some new CSS templates for a completely revamped Web site.  I found a few I like, and I’m now working on how to revise the layout to fit the new template.  No big hurry, but i may set up a second site in parallel with this one so I can start playing with the new template.

I completed the Ajax updates I had planned.  It turned out to be much easier once I got a better handle on using regular expressions in Eclipse, which is the tool I use to edit my Web pages.  That made it easy to add the span tags surrounding all of the php tags on my Web pages, and fill in the id attribute based on the tag being surrounded.  I was able to use one search and replace command instead of several tedious hours of typing by hand.  I may decide to go back and change a few things or add some more updates, but for now it works well.

I decided to finally follow through on publishing my weather data to HAM Weather  Because of their affiliations with other sites, my data also goes to PWS Weather, Weather For You, and Anything Weather.  I can see my updates on several of these sites already, although they say it can take up to 48 hours before the data shows up.  I updated the “about” page on my weather station Web site to point to those pages.  I had requested and received the documentation on how to publish my data a year or two ago.  Just never got around to it until now, as I had hardware issues to sort out first along with having a real life and day job to worry about.