Rain decimals

[Piero]

Very clear, thank you.
If that's the case, I think that adding decimal places is not very useful beyond to avoid rounding.
However, even today it continues to rain and the rounding is right.
I really think it was a startup problem, but I'll let you know in the next few days.

See you soon and thanks again

[Piero]

Today it happened again: daily rain of Cumulus: 5,4mm; WL&VWS 5,33. Same total rain, 131,1 Cumulus and 131,06 WL&VWS.

I looked in the today.ini and cumulus.009 diag files:
Start 125,6999999...
Yesterday 24,60.

WL&VWS: Start 125,73, yesterday 24,64.

Now: for Cumulus actual rain= 131,1-125,7=5.4 mm
WL&VWS: 131,06-125,73=5.33.

The problem seem to be the start value stored by Cumulus at midnight from the DLL; however, this is then compensate by the daily value, and the total rain are the same.

Really strange... what do you think?

[steve]

Sorry, I don't know. As far as I'm aware Cumulus correctly stores the current total that it's been given at midnight as the start of day total. It's hard to see how it could store something else. It then subtracts that figure from the figure that it's given during the day, to determine the current total so far today. You appear to have introduced a rounding issue somewhere by deliberately setting your rain gauge to work in different units to those that you display. I still don't understand the logic behind doing this. You still have no explanation for how Weatherlink - which you believe is displaying the correct figures while Cumulus is not - can show a value of 2.56 mm from multiples of 0.01" (0.254 mm).

Perhaps there's a bug in Cumulus. Perhaps there's a bug in the Davis DLL. Perhaps neither, I have no idea. If there's a bug in Cumulus in this area, I have no idea what the bug is. It seems to work OK for everyone else. There is a simple solution to this. Install the metric adapter in your rain gauge and make your system work in the units that you want to display.

[Piero]

I looked in the today.ini and cumulus.009 diag files:
Start 125,6999999...
Yesterday 24,60

125,699/0,254=494,878 --> is not a multiple of 0.254
24,60/0,254=96,851 --> is not a multiple of 0.254

Since Cumulus takes these values ​​from the DLL, I think that the problem is in the DLL.
So I think there are no bugs in Cumulus, the bug probably is in the conversion calculated by the DLL.

So, even increasing the number of decimal places in Cumulus, the problem would not be solved.