Perhaps others may have seen this. When I restart weather display after the computer has been off, on the console display, I get large positive spikes on the graph humidity trace, (as much as 8%) and on the temperature minus spikes of around 8 deg, and minus on the dewpoint (around 4 deg). These spikes only happen on restart. On restart of the computer,the data that is entered for the period of time that the computer is off, comes from the meteohub which runs 24/7, so I am puzzled as to why this is happening. Ideas anyone?
I’m currently using v.78 of weather display, but it was the same with v.75 and 73? The weather station is a Davis Vantage Pro and a Meteohub interface. The only time the spikes occur is with a computer restart and all other data seems to be reassuringly normal. I think? the spikes are larger the longer computer has been off. The computer is a quad core machine, 8GB memory, 1TB drive and Windows 7 Prof 64bit.
Thanks for that, no, I don’t use offsets. I thought the the offsets were purely to adjust for inaccuracies for whatever reason. I don’t know what values I would use as I thought the offsets gave a constant adjustment rather than one that dealt with sporadic events…could you explain where I should look (at least point me in the right direction!).
Thanks for getting back to me. As for your screenshot, mine is exactly the same. The only place where anything is set under offsets and limits, is under the limit tab where I have put values which avoid the sporadic running events. Strangely, when I started the computer up this morning, the spikes (both negative and positive) are much smaller. It’s also noted that these spikes only manifest themselves on the graph, not in any reading/table that I can see.
In answer to your question, if I change the period of graph time, no they don’t disappear except in the sense that they disappear off the ends as it were as the event happened before the newly displayed graph started. No it’s really weird. If the spikes had been of a constant amplitude, I might have said, oh it’s adding the values of something together, but as the amplitude varies, it isn’t that or could it be?. Maybe at some point I’ll twig it, but I’m mystified for now!
I ran into a similar problem and found the cause…not sure it’s the same as you, but worth checking. Note the time of the spike on your graph, then shutdown WD. Load the 92011lg.txt file into notepad and scroll down to the time of the spike. If there is a time slot missing, fill in the missing minute, then save the log file. Start up WD and then do a convert on the logfile, then convert last 31 days.
Here’s what my logfile looked like with the missing minute:
Yep, 2 lines with the same timestamp and different data. It would help Brian if you could capture the debug data the next time you see this happen. Under Setup->Advanced Misc Settings–>Program tab tick to log data (attachment 1). The debug data will show in View->WS2010 (attachment 2). That should show what is being imported compared to what is being plotted.
I definitely think you were on the right tack…I’ve just checked the time stamps from this morning just after 6 a.m. when I fired up the computer. There is a whole block of repeated data(different) with repeated times. This would explain why the spikes vary in amplitude. This additional data is being added to the pile and inceases or decreases the values on the graph. It would appear that this “extra data” is being compiled whilst WD is collating the missed data from when the computer was asleep. Is it WD that compiles this logfile?
I’ll look into the rest of your reply with a view to letting Brian have a “bug report”
Thanks so much for your help, it’s been invaluable.
The debug data is the text output in the WS2010 window. It will show what is being requested and sent from the station to WD when you launch WD. My guess is that Brian could use that data to troubleshoot what is happening during the import. Ideally what would happen is that after you launch WD and notice a problem, you would send Brian an email that contains the cut’n’paste text from the WS2010 window along with a snip of your 92011lg.txt file showing what happened during the import. Keep in mind that Brain is away for a week and will not be able to work the problem until he returns.
My guess, judging from all the problem posts, is that something has changed in WD recently that is affecting the Davis (and possibly other stations) import routine. I have noticed irregularities myself and I have been able to fix by hand editing the logfile.
Thanks for your comments, much appreciated. Of course the same has happened this morning when I booted the computer except that there weren’t any duplicated lines but rather some inflated figures. Interestingly, the console max/ min doesn’t seem to be affected, only the graph. I’ve included a short tract of text from the logfile so perhaps you can compare it with your own for comparisons.
I think the snip of the log coupled with the debug text should help Brian figure out what is going on. Tomorrow morning when you startup WD, capture the text from the debug output (as soon as WD finishes importing from datalogger and is collecting live data). Then post that as an attachment here along with a snip of the logfile for the same period of time so Brian can compare what was sent from the console to what was written to the logfile.
That’s what I’d intended to do. The faulty lines in the logfile are totally random. Sometimes it’s a single line with inflated/deflated figures, and sometimes it’s a block of a number of lines with the same timestamp…all good fun. Just a smally query…if I try to capture the debug data by highlighting it, as soon as new data arrives, the highlighted text flies up the screen so that I can’t capture it. Is there another way?
I have had exactly the same issues with a La Crosse machine after a restart (PC reboot) with Exit Save & exit and the normal restart of PC and then WD often ends up with a spike on baro, and negative on the humidity, and sometimes jumbled lines. I was holding back as it is a different weather station type, and Windy knew of my issue and said likely a serial port transient issue so wrapped all joints/plugs with tin foil but still get these issues at times. It only happened with the new compiler. So I have had it for a while. I have saved screen dumps and file images.