Cumulus Support

I had an interesting, very large, rain spike this morning showing up only as a "Record" high rain rate and highest hourly rain. There is also a spike on on the rain rate graph but not on the rainfall today graph. The main display correctly shows no rain today.

Does Cumulus calculate the rain rate or does that come from the fine offset console?

I did not see this reported on the current build but I could have missed the report.

Randy

Cumulus calculates all of the rain figures with Fine Offset stations; the station just supplies the total tip counter from the gauge itself and Cumulus bases everything on that. It tries to filter out temporary glitches in the counter of more than an arbitrary 30mm by waiting until it has had the same reading for six consecutive 10-second readings (so that the station will have had chance to do a 48-second update and correct itself). After this, it accepts the reading. If the counter subsequently does return to normal, you end up with a high rate being recorded but not a high total.

If the above is what happened, the diags file which was current at the time should show warnings about "large difference in rain gauge tip count", possibly followed by warnings about the counter being reset.

steve wrote:
If the above is what happened, the diags file which was current at the time should show warnings about "large difference in rain gauge tip count", possibly followed by warnings about the counter being reset.

I did not find the reset warning. Just to make sure I am looking in the right place. I reviewed the two Cumulus files in the Diag folder.

It would be in either the cumulus.008 or cumulus.009 file if it happened this morning. Your display is a bit confusing because you have the Windows default setting of hiding known file extensions, and you have .008 and .009 associated with notepad, so Windows 'knows' about those. There's no difference (other than the time they were created) between those two files and the cumulus.007 etc files.

If you zip up the diags folder and attach it, and tell me at what time the rate spike was recorded, I'll see if I can see anything unusual. Could you also include your March log file, please.

Here they are. If you do not think this is a bug don't worry about looking over the files just for my benefit.

Thanks for your time.

Randy

wd40 wrote:Here they are. If you do not think this is a bug don't worry about looking over the files just for my benefit.

I don't think it's a bug, but it's worth looking at to see whether the check mechanism should have filtered it out.

What time did the spike occur, and what value did you get for the rate?

You did ask for it in the post but time caught me and in the rush I missed posting the time.

It occurred this morning March 9th.

This is from today's high's and low's

Highest Rain Rate 33.87 in/hr 1:57am
Highest Hourly Rain 5.65 in 2:37 am

The computer was off during the night. The start of Cumulus this morning was normal except the Record light was on.

Randy

Ah, right, the code which processes the logger data is slightly different. Anyway, here's the problem:

3/9/2011 05:32:07.373 : 5:32:07 AM EWUSB: Processing history data for = 3/9/2011 1:47:03 AM
3/9/2011 05:32:07.404 : 5:32:07 AM ****Rain reset? First chance: raindaystart = 21.1417331695557, raintotal = 15.4960632324219
3/9/2011 05:32:07.435 : Writing today.ini, LastUpdateTime = 3/9/2011 1:47:03 AM raindaystart = 21.1417331695557
3/9/2011 05:32:07.435 : Creating WU URL #19
3/9/2011 05:32:07.435 : http://rtupdate.wunderground.com/weathe ... 1&rtfreq=5
3/9/2011 05:32:07.466 : 5:32:07 AM EWUSB: Processing history data for = 3/9/2011 1:57:03 AM
3/9/2011 05:32:07.544 : Warning: large increase in rain gauge tip count: 478
3/9/2011 05:32:07.794 : Writing today.ini, LastUpdateTime = 3/9/2011 1:57:03 AM raindaystart = 21.1417331695557
3/9/2011 05:32:07.794 : Creating WU URL #20
3/9/2011 05:32:07.794 :

The counter went down by 478 in one logger entry, and then back up to the correct value again. It seems to have managed to stop it appearing in the totals, but I'll have to study the code that processes the logger data to see if I can stop it affecting the rain rate also.

Pretty nifty that Cumulus has advanced to a point that it deals with these spikes with the exception of the rain rate.

Lots of good news here.

I would think this tidbit has nothing to do with the spike but I would rather give you all the information I have. I was thinking back to this morning when I started Cumulus. I can't be sure but there is a chance the wireless adapter may have lost connection during the initial wunderground update from the PC being down for the night.

Thanks for looking into this.

Randy

wd40 wrote: I can't be sure but there is a chance the wireless adapter may have lost connection during the initial wunderground update from the PC being down for the night.

I think I did see some failed 'catch up' WU updates. Unfortunately, there's no retry mechanism, but the URL is in the diags log, so you could if you wanted cut&paste those to your browser and insert your password.

I don't worry too much about the data on wunderground other than the rain so it is not a problem for me. Thanks for the suggestion just the same.

Randy