Page 1 of 1

Rainfall bug

Posted: Sat 27 Jun 2009 4:30 pm
by beteljuice
It is somewhat persisting atm :)

Rainfall last hour has suddenly zeroed.

here's data:
Rain: Last 60 min. 0.0 mm (0.00 in) | Rate: 7.8 mm/hr (0.31 in/hr)
Today: 6.8 mm (0.27 in)

Had another 6.0mm, still not updated rain last hour

Here's realtime.txt:

27/06/09 17:24:54 17.5 72 12.2 4.1 0.0 0 13.3 12.3 1016.1 --- 2 mph C hPa mm 5.7 0.0 113.5 279.6 43.4 27.4 35 17.5 -0.9 22.9 15:13 12.5 04:13 5.6 17:21 8.9 15:24 1016.5 17:18 1014.4 00:07 1.8.7 818 5.6 17.5 19.9 0 0 0 274 0.0
rain_bug.jpg

Re: Rainfall bug

Posted: Sat 27 Jun 2009 7:03 pm
by steve
Cumulus gets the last hour's rainfall directly from the station, for La Crosse stations, so I can only conclude that the station must be returning zero for that value. Has Cumulus shown the value correctly in the past?

If you turn on the debug log and leave it for 10 minutes or so, we can see the raw data that's being returned.

Re: Rainfall bug

Posted: Sat 27 Jun 2009 7:09 pm
by steve
Actually, I lied :)

It uses the 'rain last hour' from the station as the rate. It calculates the last hour's rain itself from the last hour's worth of totals that it stores. Which is possibly a bit perverse, given that the station supplies the value anyway, but it's simpler to do it the same way for all stations. I'll have a look at the code and see how it might be going wrong. The debug log will still be useful (if it's still raining) as it shows the values being added to the store of rain totals.

Re: Rainfall bug

Posted: Sat 27 Jun 2009 7:18 pm
by beteljuice
steve wrote:...
It uses the 'rain last hour' from the station as the rate. ...
Doesn't sound right - If that were the case, why can / is the rate a higher value than rain last hour ?

Yes - worked fine the last wet day - BUT, there has been lightning ?

Re: Rainfall bug

Posted: Sat 27 Jun 2009 7:23 pm
by steve
beteljuice wrote:
It uses the 'rain last hour' from the station as the rate. ...
Doesn't sound right - If that were the case, why can / is the rate a higher value than rain last hour ?
Well, yes, exactly. Isn't that the whole point of this thread? Why is your rain last hour zero when it's been raining in the last hour? It should be approximately equal to the rain rate figure.

Re: Rainfall bug

Posted: Sat 27 Jun 2009 8:02 pm
by steve
Your Jul09log.txt might be useful, too, as that will show (some of) the rain counter values that Cumulus should have been using to calculate the last hour's rain.

Re: Rainfall bug

Posted: Sat 27 Jun 2009 9:45 pm
by beteljuice
Part jun09 when rainhour had zeroed

...

27/06/09,17:07,19.8,68,13.5,1.2,6.3,15,2.4,2.4,1016.2,2461.0,27.4,34,0.0,19.8
27/06/09,17:08,19.7,68,13.4,1.2,6.3,15,3.0,3.0,1016.2,2466.7,27.4,34,0.0,19.7
27/06/09,17:09,19.7,68,13.4,1.2,6.3,15,3.0,3.0,1016.2,2466.7,27.4,34,0.0,19.7
27/06/09,17:10,19.4,67,12.9,1.3,6.3,15,3.6,3.6,1016.2,2472.9,27.4,34,0.0,19.4
27/06/09,17:11,19.4,67,12.9,0.9,4.7,22,3.6,3.6,1016.2,2472.9,27.4,34,0.0,19.4
27/06/09,17:12,19.2,69,13.1,0.8,6.3,344,3.9,4.0,1016.1,2476.6,27.4,34,6.3,17.5
27/06/09,17:13,19.2,69,13.1,1.0,6.3,292,3.9,4.0,1016.1,2476.6,27.4,34,6.3,17.5
27/06/09,17:14,19.0,70,13.2,1.5,6.3,295,5.3,5.3,1016.1,2490.0,27.4,34,5.1,17.9
27/06/09,17:15,19.0,70,13.2,2.0,6.3,300,5.3,5.3,1016.3,2490.0,27.4,34,5.1,17.9
27/06/09,17:16,18.8,71,13.2,2.6,6.3,300,6.4,5.3,1016.3,0.0,27.4,35,5.6,17.4
27/06/09,17:17,18.8,71,13.2,3.1,6.3,295,6.4,5.3,1016.4,0.0,27.4,35,5.6,17.4
27/06/09,17:18,18.4,72,13.2,3.7,6.3,291,6.4,5.3,1016.4,0.0,27.4,35,5.6,17.4
27/06/09,17:19,18.4,72,13.1,4.2,6.3,288,7.8,6.8,1016.4,14.5,27.4,35,5.6,17.0
27/06/09,17:20,18.4,72,13.1,4.8,6.3,286,7.8,6.8,1016.3,14.5,27.4,35,5.6,17.0
27/06/09,17:21,18.4,72,13.1,5.4,6.3,284,7.8,6.8,1016.3,14.5,27.4,35,5.6,17.0
27/06/09,17:22,18.4,72,13.1,5.6,6.3,282,7.8,6.8,1016.3,14.5,27.4,35,5.6,17.0
27/06/09,17:23,17.8,72,12.5,5.1,6.3,280,10.0,9.0,1016.2,36.8,27.4,35,0.0,17.8
27/06/09,17:24,17.8,72,12.5,4.5,5.6,278,10.0,9.0,1016.2,36.8,27.4,35,0.0,17.8
27/06/09,17:25,17.5,72,12.2,4.0,5.6,274,13.3,12.3,1016.1,69.4,27.4,35,0.0,17.5
27/06/09,17:26,17.5,72,12.2,3.5,5.6,270,13.3,12.3,1016.1,69.4,27.4,35,0.0,17.5
27/06/09,17:27,17.3,72,12.0,3.2,5.6,271,14.4,13.4,1016.1,80.3,27.4,35,3.1,17.3
27/06/09,17:28,17.3,72,12.0,2.9,5.6,275,14.4,13.4,1016.1,80.3,27.4,35,3.1,17.3
27/06/09,17:29,17.1,72,11.8,2.6,5.6,280,15.0,14.0,1016.1,86.5,27.4,35,0.0,17.1
27/06/09,17:30,17.1,72,11.8,2.0,5.6,287,15.0,14.0,1016.0,86.5,27.4,35,0.0,17.1
27/06/09,17:31,17.0,72,11.7,1.6,5.6,349,15.6,14.6,1016.0,92.2,27.4,35,4.7,16.0
27/06/09,17:32,17.0,72,11.7,1.4,5.6,64,15.6,14.6,1016.1,92.2,27.4,35,4.7,16.0
27/06/09,17:33,16.7,73,11.6,1.7,4.7,67,16.1,15.1,1016.0,97.4,27.4,35,0.0,16.7
27/06/09,17:34,16.7,73,11.6,1.7,4.7,67,16.1,15.1,1016.1,97.4,27.4,35,0.0,16.7
27/06/09,17:35,16.7,73,11.6,1.7,4.7,67,16.5,15.4,1016.2,101.0,27.4,35,0.0,16.7
27/06/09,17:36,16.7,73,11.6,1.7,4.7,67,16.5,15.4,1016.1,101.0,27.4,35,0.0,16.7
27/06/09,17:37,16.7,73,11.6,1.5,4.7,67,16.5,15.4,1016.0,101.0,27.4,35,0.0,16.7
27/06/09,17:38,16.5,73,11.4,1.2,4.7,67,16.8,15.7,1016.1,104.1,27.4,35,0.0,16.5
27/06/09,17:39,16.5,73,11.4,1.0,4.7,67,16.8,15.7,1016.0,104.1,27.4,35,0.0,16.5
27/06/09,17:40,16.4,73,11.4,1.0,4.7,67,17.0,16.0,1016.0,106.2,27.4,35,0.0,16.4
27/06/09,17:41,16.4,73,11.4,0.9,4.7,67,17.0,16.0,1016.0,106.2,27.4,35,0.0,16.4
27/06/09,17:42,16.4,74,11.6,0.4,4.7,68,17.1,16.1,1016.0,107.7,27.4,35,0.0,16.4

Re: Rainfall bug

Posted: Sat 27 Jun 2009 9:49 pm
by steve
Ah - your rainfall counter goes back to zero at 17:16. So when Cumulus calculated the difference after that it was negative, so it used zero. This will always happen for a short period if the counter resets. I wonder if 2500 is where the WS2300 counter wraps around?

Re: Rainfall bug

Posted: Sat 27 Jun 2009 10:32 pm
by beteljuice
Hmm .. 'twud seem so (duplicate values as I uplugged / went wireless, lost a comm port to lightning before :x )

Presumeably It will be more prevalent with mine as a 10x rain collector.

What is a 'short time', it was another hour before the rain ceased and no hour count - next day rollover ?

Re: Rainfall bug

Posted: Sun 28 Jun 2009 8:48 am
by steve
beteljuice wrote:Presumeably It will be more prevalent with mine as a 10x rain collector.
Yes, that thought occurred to me also. 2500 seems like an odd figure for wrapping around, though.
What is a 'short time', it was another hour before the rain ceased and no hour count - next day rollover ?
Well, at 1816 I'd have expected things to return to normal; that didn't happen?

Your total for yesterday is now incorrect, for the same reason, most likely.

Re: Rainfall bug

Posted: Sun 28 Jun 2009 9:15 am
by steve
I lied again :)

It uses the calculated daily total to calculate the hourly total, not the 'raw' rain counter. Since the wrap-round would have caused your daily total to be zero, the hourly total would stay zero until the rollover.

Re: Rainfall bug

Posted: Sun 28 Jun 2009 9:24 am
by beteljuice
Yesterdays Total rainfall count appears to have been maintained (16.7mm), and graphs look OK.

Doing the maths (2500 - start) + new ..... might be 0.1mm out.

Crossed post, as expected then.

Re: Rainfall bug

Posted: Sun 28 Jun 2009 10:52 am
by steve
It appears that the WS2300 stores its count of rain tips in 12 bits, so the maximum is 4095 tips, which equates to about 2121 mm (or 212.1 in your case). What's puzzling is how your counter went above that figure before wrapping. But the WS2300 is a mysterious beast...