These are the news starting 01 Jan 2013
Click here for the news of the previous years.
|16 Dec 13||Recalibration of all gas sensors. The ozone sensor shows again a fault ("relay board communication") which disappears after some tinkering. Span check using the internal O3 generator does not statbilize. O3 readings should be used with care!|
|Maintenance by T.S.E. All sensors up and running gain.|
|10 Dec 13||Repair work from 08:00 to 10:30 UTC:
1. The CO2 pump had a leaky membrane (a slit), changed membrane using a used one.
2. The main exhaust pump which sucks in ambient air from outside through a manifold going to the different sensors was noisy, what points to a near future malfunction. I changed this pump with a spare one (used, but working correctly).
3. The state of the different gas sensors is the following:
- CO2 sensor: OK
- Ground ozone sensor: bad, shut off
- NOx sensor: bad, shut off
- Zero air generator: powered on, but will probably go bad in a couple of hours. This has no further consequences...
There will be a maintenance by T.S.E. next Friday.
|01 Dec 13||Uploaded Nov13 data into data archive. Beware: there are no ground ozone data available for the full month, and neither NO2 and NO data for the second half of the month (starting 15Nov13 19:00). Sorry for this, but the ageing gaz sensors really make serious problems. Our near-zero budget does alas not allow us to replace any of these expensive instruments....|
|19 Nov 13||The NOx sensor is also down from the 16th Nov. on. So the CO" sensor is the sole remaining working gas sensor. There will be no remedial before mid-December.|
|14 Nov 13||The ground ozone sensor is definitively out of order. As the CO2 sensor pump failed again, it has been replaced by that of the non working O3 sensor.|
|01 Nov 13||Uploaded October data into the data archive. Be careful, there are quite a lot of problems during this month! The ground ozone readings should definitively be ignored: they are too low due to an instability problem that can probably not be solved before the next major maintenanceby T.S.E in December.|
|29 Oct 13||Raoul and I spent the morning looking for that mysterious "Relay Board Warning" fault of the ozone sensor.. This is an intermittent fault we were not able to pin down: all (?) voltages seem ok... the problem could be a flaky I2C bus. The ground ozone data should be considered unreliable until told otherwise!|
|28 Oct 13||The logger stopped logging yesterday...because its memory was full. This should not happen, so my face is pink! Everything restarted at 15:30 today.|
|23 Oct 13||Yet another morning spent repairing and installing
1. I replaced the noisy CO2 pump by a refurbished one (and made a mistake too silly to mention...). The "new" pump runs smoothly, and pressure and flow are ok (flow is a bit low, but still in the tolerance region).
2. The ozone sensor's pump had lost its fan --> repaired. The "relay board" failure message has vanished, and the sensor works again ok. I begin believing in miracles....
|22 Oct 13||Ignore the readings of the OZONE sensor until told that the O3 data are again ok: there now is a problem with the relay-board not communicating with the system.|
|21 Oct 13||1. Raoul shut off the Ozone sensor yesterday morning,
to avoid further damage from the stuck pump (which transfers all
electrical input into heat!)
2. I replaced the pump this morning at 07:45 UTC by a home- refurbished model which seems to run rather smoothly. Sensor is up again.
3. The CO2 pump makes disquieting noises, Will have to be replaced real soon. What a mess...
|The pump of the ozone sensor is down... it seems like the two API sensors (CO2 and O3) have reached an agreement for making maximum trouble! A repair will have to wait for a couple of days...|
|16 Oct 13||1. Installed new UPS for all relevant seven
2. The internal temperature regulation (Peltier) of the UVB sensor went bad the 12OCT13 at 17:00, for reasons unknown. I sprayed the outside connector with contact-pray, and it seems working ok again.
3. The pump of the API CO2 sensor has stopped, probably due to stuck bearings. I replaced this pump by a spare one (also an older used model), and the CO2 is ok starting 09:45. I'm running low on spare pumps (KNF PU1780, AC), a model that has been discontinued. Lets cross the thumbs, because the next breakdown will be fatal....
|1. RTL radio aired yesterday a short comment by
meteorologist Yann Logelin on the meteoLCD Iphone app written by LCD
students under the guidance of Marcel Kramer.
Look here: http://news.rtl.lu/meteo/wiederwelt/476204.html (spoken and written text in Lumbourgish)
2. If anything can go wrong, it will! Three hours after that reportage, meteoLCD went into limbo and was not accessible anymore from the web. An UPS driving the family of related computers died suddenly. It has been replaced today at 07:00 by a temporary loan; a refurbished more powerfull UPS has been ordered at www.battery-direct.de (an online specialist for UPS batteries and refurbished UPS I am very happy with since many years).
|08 Oct 13||Maintenance and calibration of NOx and ozone sensors.
Removed from rack and checked Sonimix: no detectable fault neither vibrations (problem is an shut-off of the compressor after about 5 to 10 minutes). Changed inlet filter (even if old one seems to be clean).Seems to be ok now outside rack.Reinstalled in rack: seems to be ok now delivering 1.6 bar.
|07 Oct 13||Repaired the solar radiation KIP pyranometer and put
it back to the live data page. The fault was a corroded wire inside
the pre-amplifier box. (08:30 UTC). Maintenance and calibration of
There is a problem with the Sonimix Zero Air generator. Shut off.
|01 Oct 13||Uploaded oct13.dat file into archive. The KIP pyranometer went bad starting 08SEP13 13:00 UTC. The bad data are flagged by 9999. As a valid replacement use the UVA_Bio data, which the following calibration line: KIP = 19.6279*UVA_Bio. The linear fit is quite good: R2 = 0.9834 (see here).|
by Luca Mathias and Francis Massen analysing lightning data
|21 Sep 13
||Attention: the solar radiation
pyranometer (KIP) is.down!
I have replaced the KIP data with the adjusted data from the UVA_Biometer. The relationship is:
KIP = 19.8807157*UVA with R2 = 0.9831.(21 Sep 2013, 09:55)
This must do for the moment, as a more complete analysis of the problem will be possible only in a couple of weeks.
|18 Sep 13||I have erased my old trusty Frontpage HTML editor while i nstalling Powerpoint 2007 on my laptop, so I must use the Seamonkey Composer as a replacement for a while..There are some quirks, which I ask you just to ignore with tolerant understanding....|
|meteoLCD on the air again starting 12:00 UTC. The 7days plots will be rebuilt during the coming week. I did not find the cause of the breakdown, which possibly was due to a network related malfunction.|
|meteoLCD is down since 00:30 today. Sorry, possible remedial not before Friday, 13 Sep.|
|06 Sep 13||Maintenance by T.S.E.|
|Uploaded aug13.dat file into data archive.|
|31 Aug 13||The HD failure of the Linux station played havoc with the all.dat file which holds the original unmodified logger data. I rebuilt the file all_01Nov12to26Aug13.dat from original raw downloads which corresponds to the missing data. See data archive.|
|27 Aug 13||No data from UVA sensor since yesterday...The cause is to silly to tell! I will remedy this tomorrow morning.|
|26 Aug 13|| 1.The main Linux computer went bad the 24Aug13
after 17:30. The problem comes from an HD corruption and
possibly material failure. I replaced the whole computer
with a backup model, and operation will restart at 10:00
UTC. Expect some initial problems and possible corrupted
2. Maintenance and calibration work.
|19 Aug 13||The problem of the not working temperature stabilization of the UVB sensor seems to be resolved. It was probably caused by fluid bird sh** which seeped into the connector. The spike in the UVB and UVA signals of today is caused by my work on these sensors.|
|16 Aug 13||The internal temperature regulation of the UVB sensor stopped working the 13 Aug 2013 at 18:30 UTC. Please use measurements since that day with caution!|
|08 Aug 13||The new ultrasonic anemometer went into limbo the 3rd August. I restarted the recalcitrant instrument today, which solved the problem. This is something to keep an eye on...|
|01 Aug 13||Uploaded jul13.dat file into data archive.|
|28 Jul 13||We had a very short but important wind storm around 21:30 UTC. See this picture which shows how high sudden wind brutally kills growing CO2 levels!|
|16 Jul 13||Maintenance and recalibration work.|
|10 Jul 13|| 1. The plot of windspeed/windchill has been
changed to be drawn using the new windsonic ultrasonic
anemometer readings (18:00 UTC).
2. The WBGT assembly ran dry and I did not notice it! Everything ok since today 07:30 UTC. Ignore the wet bulb readings readings at ch. 21 from 30 June to 09 July 2013. Sorry about that!
|08 Jul 13||Raoul Tholl and I dismounted the old ultrasonic anemometer (which is dead since almost 2 years, due to continuous attacks by black daws (jackdaws) which seem to have an irresistible urge to pick at and destroy the ultrasound sensors. We replaced the instrument by a new Windsonic ultrasound anemometer (from Gill Instruments Ltd, UK) which has a different design which seems to be "bird-proof". This was quite a job, with many difficulties creeping up and some acrobatic works standing on the handrail in front of the "abyss"! The plots relating to wind speed and wind direction will be modified in a couple of days, after the initial test runs.The new anemometer has two analogue 0-20ma outputs that are fed via 100 Ohm precision resistors into the datalogger.|
|01 Jul 13||Uploaded jun13.dat file into data archive.|
|14 Jun 13||Maintenance and calibration work by T.S.E.|
|01 Jun 13||Uploaded may13.dat file into data archive.|
|25 May 13||There was a bug in the 7days.dat files that remained undetected since 2002! The midnight hour line (00:00) was absent, and the preceding line (23:30) was present two times. All other files (today.dat, all.dat and the monthly .dat files) were ok. The bug caused an invisible blib in the 7days plots, and had no further consequences. Thanks to Mrs.Kramer this bug has been detected a couple of days ago, after being active during more than 11 years. The problem was in the today.pl Perl script, and it has been corrected yesterday. So if you look carefully today to the "previous days" data file (the 7days.dat), you will see a 25 May with a correct 00:00 line. Again many thanks to the sharp eyed detective!|
|Recalibration and maintenance work|
|The computer for the backup Davis VantagePro weatherstation went bonk. I managed to image the hard disk before complete failure. A replacement machine has been installed today at 13:00|
|12 May 13||The lightning machines are working again. A new system (openSuSE 12.2) was installed in the Linux server, and there were some serious teething problems. Many thanks to Gilles Massen for helping out.|
|09 May 13||Some serious problems with the meteo01 server running OpenSuse 11.1. This computer needs the ffmpeg package to make the lightning movies. I am unable to find an installable package for the old non-maintained 11.1 version. After checking out if 11.4 works (using a Live Gnome CD), I am trying an update to 11.4, and hopefully will be able to install the full package (probably from the Packman repo.)|
|29 Apr 13||I had to reinstall the OS of the meteo01 server, which shows the lightning pages and is used for other stuff. This was the occasion to begin to make the different Linux flavors of our computers more uniform. Lightning pages seem to be ok again (12:53 UTC).|
|26 Apr 13|| 1. A longer power outage created a bunch of
problems (like killing the power-supply of one of the
computers). Thx to Raoul for bringing everything back to
2. I wrote a little paper showing how to calculate the Radiation Amplification Factor (RAF) using the sudden dip in total ozone column which happened the 22-23 Apr. The RAF allows to compute the increase in UVB caused by a thinning of the ozone layer.
|23 Apr 13||I am still struggling with the new Lightning/2000 version: the author Mark Mears answered kindly several questions, but one of the most annoying problems is his choice of a Mercator projection for the radar-screen maps. With this kind of projection, the distance rings become ellipses, which looks quite funny for a small country as is Luxembourg. As a kludge, I changed the scaling on the web-page: the rings are now nearly circular, but the width of the map becomes smaller (and the letters of the text also loose their proportions).|
|21 Apr 13||The main Linux computer making all the important stuff nosed down the 18Apr13 after 05:30 UTC. The power supply ceased life (correctly diagnosed by Raoul), and destroyed physically the hard disk. I retried yesterday to replace the server with a new machine, but was unable to make an image created in Jun12 working. So the only solution was to reinstall a new Linux system with all the meteoLCD scripts and helper programs, what I did this morning. This new machine started today at 09:30. The 7days graphs will be rebuilt during the next 7 days. Probably no data have been lost, as everything is stored in a datalogger before being transferred to the computers. I guess this Sunday morning work will be considered by the Almighty as a worthy replacement for not going to the mass!|
|17 Apr 13||I installed a new version of the Lightning/2000 software driving the Boltek lightning detector. The map layout has been changed (and the station is wrongly centered at Bettendorf, not Diekirch), and I am not quite happy with this new map and its scaling. The developer of L2K has been contacted to see what can be done|
|16 Apr 13||Raoul and myself made maintenance work and recalibrations. The NO readings are possibly too low (even if the calibraion came out fine); we will keep an eye on this sensor. The NO2 readings seem ok.|
|09 Apr 13||The Luxembourg "national" radio station RTL made today in their studio an interview with me (F. Massen). You may want to listen here (be aware, Luxembourgish spoken :-))|
|05 Apr 13||Due to network related work, live data were unavailable from 03 Apr 12:30 to 05Apr 07:30. As measurement data are stored in the logger. nothing is lost!|
|03 Apr 13||New paper by F. Massen and M. Zimmer comparing the 2012 total ozone column measurements of meteoLCD and Uccle, Belgium.|
|01 Apr 13|| 1. Uploaded March data (mar13.dat) into data
2. By mistake I replaced the trends page from 14 Feb 2013 by a version missing the winter temperatures graph. This has now been fixed (with some minor updates).
|23 Mar 13||The Luxembourg radio station 100komma7 aired an interview with me (Francis Massen) on the occasion of the World Weather Watch Day. Listen here: http://www.100komma7.lu/files/2/5/234005_invite-virum-dag.mp3 (mp3). Beware, spoken language is Luxembourgish!|
|15 Mar 13||Maintenance by T.S.E.|
|03 Mar 13||Uploaded February data (feb13.dat) into data archive.|
|26 Fab 13||Maintenance and calibration work.|
|18 Feb 13||The thermal regulation UVB_Tin of the UVB_Bio sensor went down the 5th Feb 2013. As UVB is very low these weeks, the consequences are negligible. This is ma recurrent problem that apparently has noi been discovered (and fixed) by Solar Light during the last re-calibration in Sep. 2012. Powering down/up solved the problem (08:00 UTC)|
|01 Feb 13||January 2013 data uploaded into data archive|
|24 Jan 13|| 1. Maintenance and recalibration of NOx sensor
2. There was a computer problem of the Radioactivity sensor... now ok. (10:11)
|21 Jan 13||Calibration and maintenance of the O3 and CO2 sensors. New bottle with CO2 calibration gas finally was delivered by Praxair, so span calibration could again be made as it should..|
|17 Jan 13||Update of TRENDS page finished.|
|14 Jan 13||Started to update the TRENDS page to include the year 2012. This is work in progress.|
|01 Jan 13|| 1. The usual new year problem
with the 7 days graphs.. as the logger file does not
contain the year, Gnuplot gets confused the 1rst week of
every new year. This will self correct the 7th January.
Remember that the ultrasonic anemometer is still out of action (possibly for a longer time). Take wind velocity from ch. 61 (labeled Airspeed) and add an offset of 0.38 m/s to make readings of this lower mounted cup anemometer comparable to the ultrasonic anemometer at ch. 4 (labeled Air_Velo) Wind direction should be read from ch.13 (labeled Windvane); ch. 15 (labeled WindDir) corresponds to the ultrasonic anemometer, and should also be discarded.
So in short, here are the channels from left to right in the data file that should be discarded:
The main reason for keeping the data files unchanged is that a lot of users pick data from these files and rely on a stable format.
2. Uploaded dec12.dat file to data archive
and 2012.zip, which is simply the concatenated file
holding all monthly .dat files