oddball 'extreme' readings

[stevew]

still seeing some odd rainfall readings (2116mm in last hour) - note Im not updating the web atm, just watching the console here, so my initial report is that the problem persists sorry

[steve]

To be honest, I didn't expect it to make a difference. Could you turn on the debug log and let it run for while - I'm printing a message when the new code decides it needs to delay.

Thanks,
Steve

[stevew]

done

[steve]

Thanks. In case you didn't know, the debug.log file is created in the Cumulus installation folder, and gets flushed to disk every 5 minutes. I wouldn't have thought it would have to run for very long before the new code did something.

Steve

[stevew]

ok Steve, here you go.

At 13:44 All time high daily rain 2116.0mm
At 13:47 All time high wind speed 32.2mph

debug.log attached

[steve]

Thanks, that's interesting. What I hadn't realised before is just how long it takes to read all the data - about 6 or 7 seconds. This means the timings in my new code aren't suitable, it hasn't finished reading the data when the station starts to talk to the sensor. I'll adjust the timings and get back to you.

Steve

[stevew]

thanks Steve, would love for you to nail this and will be happy to try new builds and repeat the test cycle when you're ready

[bjornfred]

I have also started seeing extreme temperatures since the last release.
Suddenly it can show -30 C on the webpage.
The WS2300 shows the right temp.
http://www.anglaberga.se/vws/weather/index.htm

[steve]

The only change to the WS2300 data reading code in the latest release is the addition of the 'sync' facility, which allows you to avoid reading the data when the station is talking to the sensors - which doesn't seem to help anyway, and is turned off by default. Are you using a USB/serial adapter?

Steve

[bjornfred]

Serial, and i never seen this before the last update.

The only change to the WS2300 data reading code in the latest release is the addition of the 'sync' facility, which allows you to avoid reading the data when the station is talking to the sensors - which doesn't seem to help anyway, and is turned off by default. Are you using a USB/serial adapter?

Steve

[steve]

Is it just the temperature that's doing this? You could use the ini-file setting that causes large fluctuations in temperature to be ignored. In the [Station] section of the cumulus.ini file, add a line:

EWTempDiff=x.y

where x.y is the maximum temperature difference allowed between readings. If you're getting large spikes, you could set this accordingly (e.g. a value of 10 would be safe).

[bjornfred]

So EWTempDiff=10.10 ?

Is it just the temperature that's doing this? You could use the ini-file setting that causes large fluctuations in temperature to be ignored. In the [Station] section of the cumulus.ini file, add a line:

EWTempDiff=x.y

where x.y is the maximum temperature difference allowed between readings. If you're getting large spikes, you could set this accordingly (e.g. a value of 10 would be safe).

[steve]

So EWTempDiff=10.10 ?

Well, whatever you want, really. EWTempDiff=10.0 means that a difference of 10.0 C or more between readings will be ignored. That should eliminate the -35C you're seeing.

[bjornfred]

Is there something similar for wind and windgust ?

This is very, very wrong

Highest Gust 57,9 m/s at 18:28
Highest Speed (10 minute average) 30,5 m/s (F11) at 18:38

So EWTempDiff=10.10 ?

Well, whatever you want, really. EWTempDiff=10.0 means that a difference of 10.0 C or more between readings will be ignored. That should eliminate the -35C you're seeing.

[steve]

No, only for temperature. You can't apply the same principle to wind readings, as these could vary a lot between readings. It has some built-in fixed values for discarding obviously wrong values, but yours must fall within the thresholds (I thought I used 50 m/s, actually, so I don't understand why your 57.9 gets through). I don't really know what to suggest, normally this sort of thing occurs with serial/USB adapters. Do the readings come from the logger, or during normal running?

If you turn on the debug log, the log might give some clues, but it will probably just show that the data being read is the same as you are seeing.