Readings gone crazy - how do I correct them

[hungerdunger]

Between 13:23 and 13:24 this afternoon, Cumulus recorded the following all-time records:
Wind Speed: 217mph
Rain Rate: 3679mm
Daily Rain: 309.6mm
Hourly rain: 307.5mm

Cumulus's main display screen shows correct figures - 3mm of rain today, and average wind speed of 4.4 mph, but the "Recent Extremes" section does show the high figures listed above, and they are reflected in the graphs. The error appears to be with the equipment, as the base station is showing 306.6mm of rain in the previous hour.

I also found about 250 "Lost sensor contact" messages in the error log between 12:48 and 13:55, although none at the times the extremes above were recorded. I have been getting these messages occasionally for about two weeks, but they haven't been accompanied by bad data until today.

So, two questions: Does anyone have suggestions as to what may be going wrong with my equipment to cause this behaviour? I replaced the batteries recently and have checked them just now and they appear to be OK - are there any other checks I should carry out?

When this type of thing happens (I seem to remember odd things like this have happened before occasionally), what do I need to do to correct all my records?

[steve]

If you started Cumulus not too long before the bad data, the easiest fix is to stop Cumulus and restart from the backup before the problem (copy the files from the backup, into the data folder). This assumes that the logger doesn't actually have the bad data in it, but if the console is also showing it, then it may well be in the logger too.

Harder method if the logger has the bad data: Just restore your alltime.rec file from a backup, and edit today.ini to remove today's extremes. You will also need to edit your january log file - the format's in the help and the wiki.

All editing and file copying to be done with Cumulus stopped, of course.

[daj]

First of all check the 'backups' folder to see if you have a recent one, if not...

To correct the data, you will need to edit some files

You will not need to edit the dayfile.txt as it happened today and dayfile only gets updates at 'rollver' (midnight usually)

You will need to edit this months log file and remove the offending readings -- change the odd entries to the figures just before of after it went crazy. Edit this file with Cumulus shutdown
File spec here...http://wiki.sandaysoft.com/a/Monthly_log_files

Edit your today.ini file too, and remove the Highs!

Now, launch Cumulus and click Edit, All time records. Chances are you will not remember what the previous high was, if you do type it in. If not, you will need to go to a backup alltime.rec file from the backups folder - try a recent one. Close down Cumulus, copy the backup alltime.rec and replace the one in the data folder

Start Cumulus backup

[hungerdunger]

Thanks Steve and daj for your advice. Luckily a simple restart using the last backup sorted the problem, but at least I now have your instructions on what to do next time if that doesn't work.

As to my station's bad behaviour - erratic data and losing contact with the sensor - does anyone have suggestions on what to look for. It's just over a year old, and although the manual doesn't mention maintenance, I wonder whether I should be doing some.

[steve]

Mine is very bad at losing sensor contact, and it's not very far away - maybe 30 feet. It's through two breezeblock walls, though.

[hungerdunger]

Mine is very bad at losing sensor contact, and it's not very far away - maybe 30 feet. It's through two breezeblock walls, though.

Mine is more like 60 foot and through an 18" stone wall (although the console and sensor can almost "see" each other through the window). What concerns me is that the "Error" light has hardly ever come on until about two weeks ago, and now it's every couple of days.

[beteljuice]

You say 'no maintenance' - the combination of the 'old' batteries and the cold temps may be enough to reduce the range of the transmitter.

I know with the LaCrosse 2300, if working cabled, the feed voltage is higher than the batteries, although there will be some loss through the cable.

Has anyone been brave enough to try an additional 1.5v / 1.2v battery cell, or power from a (typical) unregulated power supply ?