Page 1 of 1

Invented rainfall since start of 2018

Posted: Thu Jan 04, 2018 2:45 pm
by oh6hps
I started to get invented rainfall (30-40 mm) at the beginning of this year. It occurs one or two times/day without actual rain or related values on VP2 console or Meteobridge. They both show reasonable values. I have read a bunch of threads/FAQ but I'm still a bit lost with my issue. It is not a biggie, but it would be nice to find out the root cause of it. I can live with that by editing Todays rain of MX (done that couple of times).

Any help where to start from would be appreciated. I didn't find any errors from MXDiags logs.

Setup:

Davis VP2 + IPLogger (latest fw´s) -->Meteobridge --> Raspberry Pi 2 (mono 4.0.5) Cumulus MX v3.0.0 (3043)

Re: Invented rainfall since start of 2018

Posted: Thu Jan 04, 2018 3:35 pm
by steve
Cumulus uses the annual rainfall total to determine rainfall amounts, so it looks like something odd is happening to your counter since the start of the year. The diags file(s) will have more information about the value of the counter.

Re: Invented rainfall since start of 2018

Posted: Thu Jan 04, 2018 3:50 pm
by oh6hps
Thanks Steve! There is no error doesn't mean there isn't something interesting...

Here is my log from this morning on. I also turned debug logging on while ago. There is some lines about raincounter reset and setting. If you have time to take a look.

Re: Invented rainfall since start of 2018

Posted: Thu Jan 04, 2018 3:59 pm
by steve
Your annual rain total as provided by the console (or from Meteobridge?) apparently keeps flipping between 40.2 and zero. Analysis of the data packets after you turned on the data logging would confirm this (the annual counter is in there somewhere) but I’m not in a position to do that at the moment.

Re: Invented rainfall since start of 2018

Posted: Thu Jan 04, 2018 4:06 pm
by oh6hps
Ok, I leave the data logging on for next flipping. Just Data logging or both debug+data? Many thanks again!

Re: Invented rainfall since start of 2018

Posted: Thu Jan 04, 2018 4:09 pm
by steve
Both.

Re: Invented rainfall since start of 2018

Posted: Thu Jan 04, 2018 4:23 pm
by oh6hps
Both are on now. Meteobridge (uptime 12 days) show total rain 37.2mm and Davis console annual rain 40.4mm.

Re: Invented rainfall since start of 2018

Posted: Fri Jan 05, 2018 7:28 pm
by oh6hps
Steve says something and computer says: Yes. Suddenly everything seems to be ok for now. Still under monitoring/logging, though.

Re: Invented rainfall since start of 2018

Posted: Sat Jan 06, 2018 8:14 am
by oh6hps
I was jumping too early, it happened again at 2018-01-06 04:38:59.191. The MXDiags log attached since midnight.

Re: Invented rainfall since start of 2018

Posted: Sat Jan 06, 2018 8:07 pm
by steve
MX works by reading 50 'LOOP' packets and processing the data, and then reading a single 'LOOP2' packet for a few extra bits of information. What is happening is that when MX asks for the 50 LOOP packets, sometimes Meteobridge is apparently sending a LOOP2 packet again rather than a LOOP packet. MX doesn't check the type of packet it's been given (it should, really) so incorrectly interprets these LOOP2 packets as LOOP packets.

This means its getting incorrect values for things like the annual rain total (it reads it as zero). If it just happens once, MX ignores it. But sometimes it happens again the next time MX asks for 50 LOOP packets, and having had two consecutive zero rain counter readings, MX decides that the rain counter really has been reset to zero. When it then gets a LOOP packet with the correct total, it applies the new figure, resulting in the sudden jump in the total.

I can change the code so that it checks the packet type (but I have no idea when I will be able to release a new version) but the basic problem appears to be with Meteobridge sending the wrong packet or something causing the LOOP2 packet to be repeated.

Re: Invented rainfall since start of 2018

Posted: Sat Jan 06, 2018 8:22 pm
by oh6hps
Ok, thank you so much Steve for your comprehensive answer. I’ll ask if there is some workaround on Meteobridge end.