Page 1 of 1

UV Spike removal

Posted: Thu 07 Apr 2016 11:42 am
by AndyT
Hi all

Does anyone know if it's possible to implement a spike removal for UV? It doesn't happen often but occasionally I get a UV 12 or 13

Can it be done via the Cumulus ini?

Thanks

Andy

Re: UV Spike removal

Posted: Thu 07 Apr 2016 11:52 am
by steve
I can't look at the code right now, but I don't recall doing any UV spike removal.

Re: UV Spike removal

Posted: Fri 08 Apr 2016 4:37 pm
by AndyT
Thanks Steve,

I'm not sure that it would work anyway, after today! We had a lot of cloud about and of course when a bit of blue sky appeared, the UV shot up instantly.... from like 3 to 7....

Andy

Re: UV Spike removal

Posted: Fri 08 Apr 2016 4:47 pm
by steve
It does sound like it would be tricky to cater for.

Now that I'm home again, I can confirm that there's no UV spike removal in Cumulus.

Re: UV Spike removal

Posted: Sat 09 Apr 2016 12:44 pm
by AndyT
Thanks Steve, probably best left alone, this one :)

Re: UV Spike removal

Posted: Tue 19 Apr 2016 6:42 pm
by philpugh
Sorry to butt into this topic but...

I've just installed a second weather station - a WS3083 - and whilst the Solar seems to work OK - it matches the MaxSolar values on clear sunny days - I noticed a weird value last week. CumulusMX had recorded a max value for UV of 16 at 4:06 am. Unless I missed a very large supernova - I suspect this is a real oddball. The interesting thing is that it was displayed on the :8998 port web page but wasn't recorded in the Realtime database values - it appeared in the Dayfile db and dayfile.txt. It hasn't happened again and as the unit was 9 meters off the ground It is unlikely to be a spurious light source - I'm out in the sticks - no street lights (or mobile phone signal or anything above 2.5Mbps brodband) so I'm at a bit of a loss. FWIW I suspect the UV sensor on the WS3083 is a bit sensitive and may be responding to the top end of the visible spectrum. I've edited it out of the records!

Edited to correct file names

Re: UV Spike removal

Posted: Tue 19 Apr 2016 7:00 pm
by steve
philpugh wrote: CumulusMX had recorded a max value for UV of 16 at 4:06 am.
Cumulus treats any value of more than 16 as 16, so it could have been some kind of data corruption.
The interesting thing is that it was displayed on the :8998 port web page but wasn't recorded in the Realtime or Monthly database values
It may have been overwritten by a valid reading before the next realtime update.
it appeared in the Dayfile and alltime.ini.
It would have appeared as the day's high in dayfile.txt, but UV doesn't appear anywhere in alltime.ini

Re: UV Spike removal

Posted: Tue 19 Apr 2016 7:04 pm
by philpugh
Steve - you beat me to it - I've just edited the OP to correct the error - indeed it did appear in the dayfile.txt - not alltime.ini (or alltime.txt) as these don't record that information. Sorry.

Yep - treating it as an oddball value - will keep eye out for it happening again.