Page 1 of 1

Another CWOP problem

Posted: Thu 15 Jul 2010 2:15 am
by AK7V
Hello,

I am trying to submit CWOP weather data using Cumulus directly instead of using the wxnow.txt file and another APRS program. Problem is, Cumulus seems to be putting some garbage in the weather packet. Here's an example:

2010-07-14 18:32:20 PDT: AK7V>APRS,TCPIP*,qAC,FIRST:@150131z3421.58N/11739.14W_198/001g005t079r000P18446744073709551608h38b10234eCumulusDsVP

What are all those numbers after P? It's making me record some weird precipitation. Also, I noticed that Cumulus is sending something incorrect to Weather Underground (rapid fire) for rain. It says "N/A" for rain in the rapid fire updates.

Here's a link to my Weather Underground: http://www.wunderground.com/weatherstat ... =KCAWRIGH6

And here's my FindU APRS/CWOP data: http://www.findu.com/cgi-bin/raw.cgi?call=AK7V

Thanks,
Jason

Re: Another CWOP problem

Posted: Thu 15 Jul 2010 7:04 am
by steve
The two problems are related; the "P" figure for CWOP is 'rain today' in tenths of an inch. I've looked at your CWOP data and WU data and they both look OK now - well, sending zero rather than that huge figure.

I don't know how your 'rain today' figure got corrupted; it's not normally a problem with Davis stations. I'm guessing that when you started it up it somehow got an incorrect figure for the start of day counter. If you zip up the diags folder and attach it, I'll see if I can see anything.

Re: Another CWOP problem

Posted: Thu 15 Jul 2010 1:12 pm
by AK7V
Thanks Steve. Yep, it did fix itself today. And I think I figured out what happened -- somehow, during my weather-software-switching blitz, -0.08 inches of rain got recorded. Negative rain! So that was in the record for rainfall "today," yesterday, and messed things up. Now that it's the day after, it's gone.

-Jason

Re: Another CWOP problem

Posted: Thu 15 Jul 2010 1:22 pm
by steve
AK7V wrote:somehow, during my weather-software-switching blitz, -0.08 inches of rain got recorded. Negative rain!
Ah, right, and the Delphi "Format" function which I use for generating the string clearly gets confused when asked to convert a negative number into an unsigned decimal and includes the sign bit as part of the number.