M5 Maintenance Log

I’ll be using this thread to store information about maintenance to the NWTC M5 (Site 4.0) met tower.

You can subscribe to an RSS feed of this topic.

[size=150]November 12-15, 2013: DAS failure[/size]
Following an update to MS Windows, the M5 DAS failed. The system was upgraded to Windows 7 / Labview 2012 and the DAS was was operational again by Friday 15th November.

November 18, approx 18:00 UTC: 100-m sonic settings updated to fix a problem with intermittent ramps.

November 30, 11:20 UTC - December 2, 22:10 UTC. Hard drive failure leading to loss of data. Not detected until Monday 2.12 because of Thanksgiving holiday. Hard drive replaced and system restarted.

December 19-24: A change to the data acquisition system (DAS) coding caused the DAS to slow down so that we were not getting true, 20 Hz data acquisition. Automated QC routines caught this error and flagged it in the data files. The change to the DAS coding has been reversed and we are now getting 20 Hz data again. This does however mean that there are approximately 4 days of data that have been lost.

January 5, 2014.

A problem with file permissions between January 1 and January 6 meant that although data were being collected and stored by the DAS, they were not being processed. That bug has been fixed and the files are now being reprocessed. I expect the files for this period to be completely processed by about 23:00 UTC on January 6, 2014.

The 74 m and 61 m sonic anemometers on the M5 tower were replaced at on Friday 14th March 2014. Data are available from approximately 17:00 UTC. There were some outages around this time as the sonics and DAS were adjusted.

The M5 tower is out of service this week (June 23-27, 2014) to allow instruments on the tower to be swapped out. We do this every 12 months to keep devices in calibration. We anticipate data will be available again from June 27, possibly June 30, 2014.

The 3m dewpoint sensor was replaced around 15:00 local (22:00 UTC) on August 6, 2014. We have changed from the original lithium-chloride coated sensor to a solid-state sensor from Vaisala. We have noticed a few events where the signal from this sensor has not been steady, and caution that the dew point temperature may be unreliable. This effects the Richardson number as well, as the virtual temperature gradient is possibly incorrect. Data derived from this measurement will be flagged as possibly bad in the data files, in the next few days.

The M5 DAS was taken offline on November 13 to trial new DAS software which is expected to substantially improve reliability. We expect to have the new DAS running by early January 2015. The raw data format may change at this time, which will necessitate some rewriting of the data handling code. This may delay delivery of raw and processed data to the web archives.

The new M5 DAS went live on January 1 and we have completed in-house testing. The new DAS has the following features:

  • Raw data are written using an XML-based format, called labview TDMS. we do all the conversion for you, so you will see no change to the raw data on the web server.
  • Sonic anemometers run at 20 Hz. All other sensors run at 1 Hz, but data are still written to file at 20 Hz. You will see this in the raw data on the server. You will have to downsample the 20-Hz data in the file to 1 Hz before calculating statistics.

New M5 data are now streaming to the archives and being displayed on the main M5 page. The new data is version 1.4. There should not be any old data available.

June 8, 2015: Annual Maintenance

M5 systems will be down for annual maintenance for 1-2 weeks from June 8, 2015. During this time we will be swapping in new / recalibrate instrumentation and updating the DAS computer.