Sandaysoft

Hi Steve

Firstly I'm running v1.85, build 690, WinXP with all the latest updates under an Admin account. My weather station is a LaCrosse WS2315.

Since the 26 Jan my dayfile.txt and yesterday.ini haven't been updating and when I boot up the computer first thing in the morning I'm getting some strange spikes on most of the graphs. See http://www.sedgleymanor.com/wx/trends.htm around 6am when the temperature appears to have jumped from 20C up to 34C when the actual temp was 20C throughout this period.

These spikes/errant data doesn't appear when I use the HeavyWeather software that came with the station.

_________________
Website : http://www.sedgleymanor.com/wx
on the Indian Ocean coast of Western Australia (down the bottom left-hand corner of Oz)

Are you using a serial/USB adapter?

Steve

Hi Steve

Serial, but I've been running Cumulus since last November with little or no problems.

_________________
Website : http://www.sedgleymanor.com/wx
on the Indian Ocean coast of Western Australia (down the bottom left-hand corner of Oz)

You mean a direct serial connection, from the station to the PC - no USB involved?

Can you attach the file from the Diags folder from the start up this morning, please?

Steve

I've attached cumulus.008.txt (renamed from cumulus.008 as it wouldn't accept the 008 extension)

I use Wireless from the Sensor unit outside to the Base Station near my computer then serial cable to COM1

Thanks

Ian

_________________
Website : http://www.sedgleymanor.com/wx
on the Indian Ocean coast of Western Australia (down the bottom left-hand corner of Oz)

I changed the list of allowed extensions to include .001 to .009, I wonder why it didn't work...



Ah, good.

The reason for the lack of rollover is that you have your station set to log at 1-minute intervals, so at start up in the morning, there is only about 3 hours worth of data in the logger, so not enough to cover the period back to the previous night. Cumulus doesn't do the rollover unless it has the data for the rollover period.

I can't explain why the temperature spike occurred - that appears to be the value that the station gave. Which other items had spikes? I can't get your web site to open.

Steve

Hello again Steve

Many thanks for quick reply. Ive increased the log time to 10mins, hope that cures the problem.

Unfortunately the server that my website is on is off-line at the moment (doesn't seem to be my lucky day does it? ) but I've attached the current graphs of my weather station. As you can see the Inside Temp, Outside Temp, Pressure and Humidity are all giving stange readings.

All the spikes appear when Cumulus is started after a period of not running it. Any idea why the yesterday.ini and dayfile.txt aren't updating?

_________________
Website : http://www.sedgleymanor.com/wx
on the Indian Ocean coast of Western Australia (down the bottom left-hand corner of Oz)

It's significant that the odd values aren't true 'spikes' (i.e. one-off garbage data readings), in that they gradually change to the new values and then go back again - from what I can make out. This looks more like some strange sensor problem, but I can't explain why more than one sensor is apparently affected. It will be interesting to see what the graphs look like now you've changed the logger interval, and the times you don't have data for aren't confusing things.



Yes, sorry, I should have been more explicit. That's what I meant by 'lack of rollover'. Those files are updated during the rollover which you aren't getting because you don't have the data to cover that period.

Steve

Hi Steve

Sorry to be a nuisance, but I'm still getting strange readings when Cumulus is started up. I ran the program all through the night and the dayfile & yesterday files updated OK, the Feb09log file also updated every 10mins as it was supposed to do.

However when I shut down Cumulus at just after 0840hrs this morning and then restarted it just after 0850hrs I got the same short of "spike" again. As you can see from the Feb09log there are now readings for every minute between 0840 and 0850hrs.

Any ideas???

_________________
Website : http://www.sedgleymanor.com/wx
on the Indian Ocean coast of Western Australia (down the bottom left-hand corner of Oz)

Your station is still logging at 1-minute intervals. Whether this is related to the spike problem or not, I don't know. But unlikely, I'd have thought. It does mean that you won't get a rollover if you don't run Cumulus overnight. I'll decipher the data in the files you've uploaded and see if I can see anything.

Steve

Hi Steve

My Cumulus is definitely set for a 10-minute Data log interval (see attached ini file) and it works perfectly while the program is running. It only goes haywire when it starts up.

_________________
Website : http://www.sedgleymanor.com/wx
on the Indian Ocean coast of Western Australia (down the bottom left-hand corner of Oz)

Not Cumulus, the station. It affects how long a period the station can hold in its memory, it only has about 175 entries.

Steve

I think I see what's happening. The data from the logger that Cumulus thinks is for 1-minute intervals from 0841 to 0849, is actually the data for 10-minute intervals from 0720 to 0840, so there's a sort of 'repeat', causing the glitch. It's as if Cumulus thinks your station is logging at 1-minute intervals, but it's actually logging at 10-minute intervals, so Cumulus is assigning the wrong timestamps to the data (the entries don't have timestamps, Cumulus has to calculate them based on the logging interval). Follow that?

Is there some way for you to find out what interval the station is actually set to?

Steve

I can set what Heavyweather calls the "History Period" in their software, which I THINK must write back to the WS2300, to anything from 1 minute to 60 hours. Because the station can store 175 sets of data this gives a total data storage time of between 175 minutes to over a year.

I had both the Heavweather and Cumulus set at 10mins. I wonder if that's what causing the problem?

I've now changed HW to 1 hour (which is the default) so I'll see what happens when I boot up in the morning.

Good night and many thanks for your patience.

Ian

_________________
Website : http://www.sedgleymanor.com/wx
on the Indian Ocean coast of Western Australia (down the bottom left-hand corner of Oz)

If anything, having the station and Cumulus at the same interval is probably a good idea rather than not. But it will be interesting to see what happens now you have it set to an hour.

Steve