GW1000 - Temp spikes

[Big Daddy]

Do the removals work on the Ecowitt GW1000.

I too had 2 x instances of high temp within 5 minutes. I have spike removal set to 5C on outside temperature. I log every 1 minutes and data below is from May2020log. Dont see anything in MX diags around that time apart from recording new highs.

Date (dd/mm/yy) Time Temperature
21/05/2020 09:43 19.5
21/05/2020 09:44 19.5
21/05/2020 09:45 60
21/05/2020 09:46 19.2
21/05/2020 09:47 19.1
21/05/2020 09:48 60
21/05/2020 09:49 19.1
21/05/2020 09:50 19.1

I think data is sent from the WS80 sensors every 4.5s so sont know if this is sampled at that rate by Cumulus. Funnily it doesnt show up the graphs for my station on ecowitt.net

Any ideas?

Thanks

[mcrossley]

No spike removal does not apply to Ecowitt stations - it only applies to the original Fine Offset stations.
Cumulus reads the GW1000 data every 10 seconds.

If this occurs again I'd appreciate a log file with debug and data logging enabled please. Lets see if the spike is in the data being supplied, or something going wrong with the decoding of it.

[Big Daddy]

Thanks Mark,
Does this mean I need t have debug and data logging permanently enabled for the time being to capture the data you need.

If so, I suspect this could generate a massive MXDiag file. Should I be ok to clear this out everyday or should I retsart MX on a daily basis to generate a new one. Obviously dont know when its likely to happen next, if at all and if it happens, how long after the even would I notice it.

I have turned them on now and one thing I also noticed was this message on every data read.

Unknown sensor id found = 107, at position = 94

Andy

[mcrossley]

Yes, you will have to leave it running with those options enabled. Massive - well in an age when we stream films and download DVDs for fun, not that big really

The error has been introduced by the latest GW1000 firmware update, they have added the WH34 battery status to all the messages sent. It's benign as the extra data is at the end of the packet. The next release fully supports the WH34 devices and is just about ready to go.

[Big Daddy]

OK Mark, its running. Will keep my eye on the size and stop / restart if it starts to get too big. I have about 9Gb of free space on my Pi SD card so we should be good for a while.

Andy

[Big Daddy]

Hi Mark,
So today (8th June)I got another spike. This time on rainfall.

Looking at the MX graphs, around 15:22 suddenly 15.9mm of rain appeared. No indication of any rain rate on graphs or in records.."Today" records show 15.9mm of rain with a High Hourly rainfall of 15.9mm @ 15:58. Looking at the MX graphs, rain falls to 0mm @ around 17:11 and then back up to 15.9mm @17:13.

Graphs can be seen at http://midlifedad.me.uk/weather/trends.php and click on the Rainfall button under the graph.

Ecowitt.net does not show any rainfall today, nor does the Live data on WS view.

Can confirm it didnt rain today.

As previously suggested I had data logging and debug logging turned on and have attached MXdiag files.

Let me know if you need anything else.

Ecowitt GW1000 is running 1.5.9.

Andy

[mcrossley]

Thanks, I'll take a look at this tomorrow...

[mcrossley]

Right - I see the problem.

At 14:57 your rain gauge went offline. MX then treats that incorrectly as a zero value for the rainfall - after three consecutive lower readings, it resets its rain counter to the zero value. Then at 15:21 the missing values reappeared, with the rain counter now back at 15.9, so MX added 15.9mm to the days rain.

Same 17:11, some sensors including the rainfall went offline, then came back again.

I'll get this fixed in code, meantime could you look at improving the reception from that sensor which would stop this happening?

[Big Daddy]

Thanks Mark for the quick response.

I will take a look at repositioning the rain gauge as it only shows 50% wi-wi signal on the WS View app.

Potential schoolboy error.
The rain gauge is located at the far side of my shed roof but there is a 20ft x 2" steel pole with 3 x catenary cables that supports the anenometer between it and the GW1000. Its only about 20m away from the GW1000 with an external wall in between. Will look at bringing it the to the other side of the shed.

Andy

[Big Daddy]

Have relocated the rain gauge and also repositioned the GW1000 so now have a "full house" of wi-fi signal on all my sensors.

Sensors.png

[mcrossley]

I'll get a refresh of the code to prevent this soon.

[rogerthn]

Temperature spikes today around 11
Station type Ecowitt GW1000 (GW1003)

Started using

Code: Select all

sudo mono CumulusMX.exe -debug

Debug logging and Data logging is enabled.
100k in 10 minutes
I do have a USB drive connected to my Raspberry

Code: Select all

Filesystem      Size  Used Avail Use% Mounted on
/dev/root       916G  3.6G  866G   1% /

[mcrossley]

Temperature spike removal only works on traditional FO USB stations.

[rogerthn]

Temperature spike removal only works on traditional FO USB stations.

Thanks!

[mcrossley]

But can you post the log file anyway - I'd like to just check the raw data coming from the GW-1000

Did the Ecowitt software/web site also record the spike?