Trends >> tempm graph

[wa7hyd]

It appears that the "tempm" graphs left hand legend is in millibars rather than degrees.

[gemini06720]

Indeed.

It appears that the values for the left hand legend are wrong - I just compared the graphic with the weather data collected by my 'other' weather software (that has been running non-stop for the past 5 years) and none of the maximum temperatures match those displayed on the graphic image...

[steve]

It appears that the "tempm" graphs left hand legend is in millibars rather than degrees.

No, definitely in degrees. You probably have a 'default' entry in dayfile.txt which is throwing the whole thing out, possibly from the first day you ran Cumulus (but they can unfortunately occasionally occur for other reasons). Have a look in dayfile.txt and see if you can see it; it should be obvious as it will have ridiculous figures in it for min and max temperature (999 and -500, if I remember correctly).

If the above doesn't account for the problem, then a picture paints a thousand words, as they say - could you upload a screenshot of your graph (or supply a link if it's on the web somewhere). And attach your dayfile.txt, please.

[steve]

It appears that the values for the left hand legend are wrong - I just compared the graphic with the weather data collected by my 'other' weather software (that has been running non-stop for the past 5 years) and none of the maximum temperatures match those displayed on the graphic image...

The min/max/avg temperature graph is drawn directly from the figures in dayfile.txt, so if your graph is wrong then there is something fundamentally wrong there. I suppose the first thing to do would be to confirm that the figures on the graph match those in the dayfile, and if they do, then we need to see why your figures don't match your 'other' software. If you have Cumulus running 24/7, they should match exactly, otherwise there may be small discrepancies (Cumulus 1 only uses the 'current' temp from the VP archive records, and not the highest and lowest over the archive period).

[wa7hyd]

Four the error entries, about six days worth, averaged corrections to fill in the bad time and it works fine now.

Thanks to all who pitched in with comments and solutions.

I also changed the Trendst file to display all large charts, better for viewing and printing when needed, which I'll have to monitor closely when updates are issued.

[steve]

Four the error entries, about six days worth, averaged corrections to fill in the bad time and it works fine now.

If it happens again, could you please send me the contents of the diags folder, so I can see why it's happening?