Data spikes with Fine Offset (EasyWeather WS1090)

[jschauer]

I have been working with a client to troubleshoot data errors with the Easy Weather WS-1090 wireless weather station. I have Easy Weather 6.2 and Cumulus 1.9.4 running on WinXP. We recently moved the station to a new computer. I deleted all historical data and started it fresh. The trends graph and records began to show spikes, with wind speeds gusting to 129mph. Temps and pressure readings also spiked. I read documentation, wiki, and searched the forum. I suspected that the station was not synchronizing correctly, so I downloaded and ran the SetLogger utility, setting the update interval to 5 minutes. I cleared the data directory, restarted cumulus. Spikes continued to appear regularity. I then tried entering values in the calibration configuration page, entering spike values to be ignored. I again cleared the data directory, and restarted Cumulus. The trends graph and max/min values continue to be recorded; however, the magnitude of anomalous readings have been reduced significantly. (Winds have been pretty calm, or else the spike value of 15m/s we entered is causing all wind data to be ignored.

The weather station touchscreen appears to be reading correct data continuously. The sensors seem to be communicating over the wireless connection to the station. Communication between Cumulus and the website over FTP are working. Problems seem to be between the Cumulus software and the weather station over the USB connection.

The website trends page can be viewed at http://goldstreamsports.com/weather/trends.htm . The computer is located at the client's shop, which is a considerable drive from my location. I do not have remote access to the client's computer, I do have FTP access to the web server files. I would like to get some ideas for further troubleshooting before I return to work on the client's computer. I have tried talking them through some trouble shooting steps, but I need to be there to access local files or capture screen shots.

Thanks in advance for any help you can offer.
John Schauer

[steve]

'Spikes' are a well known problem with Fine Offset stations. Cumulus automatically throws away the more obvious of the garbage data, and most people successfully use the 'spike removal' settings on the calibration screen to filter out the rest.

I believe some people have reduced the problem somewhat by using a USB cable with ferrite beads.

Those graphs do look rather odd, though, and perhaps you have some other problem. If you zip up the diags folder and attach it, I'll take a look. You're not trying to run Cumulus and Easyweather at the same time, are you? Or have accidentally started two copies of Cumulus running? Have you selected the correct station type (i.e with or without solar sensors?).

[jschauer]

Thanks, Steve. I think I have the station type selected correctly with the correct sensors. We have had both EasyWeather and Cumulus software running. I'll get them to turn off Easy Weather and restart. Next time I'm out there, I'll get the diags folder, zip it, and send it to you.

John

[steve]

If you have Easyweather and Cumulus running at the same time, that's almost certainly the cause of your problem. Two programs can't communicate with the station at the same time.

[jschauer]

Thanks! That is an easy solution! I hope you are proven correct!

[Cambium]

'Spikes' are a well known problem with Fine Offset stations. Cumulus automatically throws away the more obvious of the garbage data, and most people successfully use the 'spike removal' settings on the calibration screen to filter out the rest.

Quick question regarding this topic..

My pressure right now is reading 1044mb but actual is closer to 1027mb. Regarding the spike tool, what do people typically set it at and when does it calculate the spike?

In other words, if I set the spike removal at 10.0mb will it ignore a 10mb spike in a 5 minute time frame, that minute, 1 hour, ect?

In fact, my reading doesn't even look like a spike. Not sure why it's reading high. Temps are cold out there, could the cold air be affecting the sensor?

[steve]

Spike removal works by comparing successive readings from the station.

Your reading might be due to a faulty sensor, but it's probably because you need to set the relative pressure on the station. Instructions are in the manual, and there's also some information in the FAQ: http://wiki.sandaysoft.com/a/FAQ#Cumulu ... et_console

[Cambium]

Spike removal works by comparing successive readings from the station.

Your reading might be due to a faulty sensor, but it's probably because you need to set the relative pressure on the station. Instructions are in the manual, and there's also some information in the FAQ: http://wiki.sandaysoft.com/a/FAQ#Cumulu ... et_console

Thanks for the quick reply! I just realized the station itself says 1032 but Cumulus says 1044. Hmmm. Looking for the adjustment area but this is a first time its happening.

[Cambium]

Nevermind! Calibration set to 12 on the pressure LOL