Problem with windspeed graph

[steve]

The earliest back-up log was at 12:48pm today, well after the overnight data was already downloaded.
I can attach something here for you tomorrow, when I download the night's data using the new Cumulus version you have supplied me.

Ah, right, you've been restarting Cumulus quite a bit, then? OK, tomorrow it is, then.

[DaveNZ]

Here's the file in diags after it got the overnight data this morning around 9am (it drew the graphs strange again).

cumulus.008

[steve]

It's giving the 'error' value (-32768) for wind direction:

2009-12-11 09:00:42 : VP2: processing logger data entry 0 for 10/12/2009 11:22:00 p.m.
2009-12-11 09:00:42 : ih=56.0 oh=91.0 it=24.9 ot=15.6 dp=13.4 pr=1017.82 wg=0.0 ws=0.0 wd=-32768 rt=0.0 rr=0.0

Cumulus handles all three of the wind parameters in one go, and ignores them all if any are reported as 'error', so this is a bit of a pain. The error value is supposed to be reported when there is a problem with the sensor. This might sound like a silly question, but is it definitely absolutely calm at the time these values are reported? It just seems rather odd to me that every night while your PC is switched off, you get mostly zero wind, and all day you hardly get any periods of zero wind. But then again, I know nothing about the weather in NZ!

[DaveNZ]

Yes there's definitely been periods of calm overnight - currently got a nice high pressure system to thank for that. Not many periods of zero wind during the day since I'm only about 2km from coastline

[Super-T]

Steve
You can compare Dave's data with mine as he is only 500 metres away down the hill a bit.

[steve]

It's just odd that no-one else seems to have ever seen this problem (including me). Of course, simple bugs can and do go un-noticed for a long time. It would just be good to have confirmation that this always happens (the error code for direction when there's no wind), then I can simply cater for it. It would also be handy to know if the same value is reported during live running, but I'd have to put more tracing in for that.

Unfortunately we didn't quite get to zero wind here, so I can't compare. I guess I'm going to have to get the ladder out and stop my anemometer, in the interests of science

[DaveNZ]

It accepts the 0 values while the data is feeding to cumulus live from the station. I have attached a small graph from a period overnight, notice the lines drop to zero before 11pm (I closed cumulus around 11.20pm). Seems that once the program is closed, after the next non-zero speed on the logger all zero values are treated as errors.

windgraph1.PNG

Unfortunately we didn't quite get to zero wind here, so I can't compare. I guess I'm going to have to get the ladder out and stop my anemometer, in the interests of science

Hehe will be interesting

[steve]

It accepts the 0 values while the data is feeding to cumulus live from the station. I have attached a small graph from a period overnight, notice the lines drop to zero before 11pm (I closed cumulus around 11.20pm). Seems that once the program is closed, after the next non-zero speed on the logger all zero values are treated as errors.

It's not that the zeroes are treated as errors, it's that the station (or the Davis DLL) is actually sending the error value for wind direction when it's calm. So Cumulus is ignoring the wind data. But you've now confirmed that it doesn't do that during live running. And we were actually becalmed briefly here overnight, and I made sure to turn Cumulus off overnight, and I got the same error values this morning from the logger.

I have no idea why it sends 'error' for 'calm' (or why it doesn't do the same when live) but the fact is that it does, and I'll have to work around it.

[DaveNZ]

I have no idea why it sends 'error' for 'calm' (or why it doesn't do the same when live) but the fact is that it does, and I'll have to work around it.

Ok cheers hope you can get around it as the wind often drops to calm here in the middle of the night. Good luck

[beteljuice]

... it's that the station (or the Davis DLL) is actually sending the error value for wind direction when it's calm...

According to WL blurb (which of course may not be the same as what the dll puts out)

As the sample rate changes with wind activity, it is possible to have a Hiwind (speed) value but NO direction.

Just to increase your headache

[steve]

(which of course may not be the same as what the dll puts out)

It's impossible to say; the DLL documention is poor and has errors anyway.

[steve]

New build now available.

[DaveNZ]

New build now available.

Wow thank you Steve that was quick.
Wind stayed above zero here all night so graphs are good this morning. Will install the new build and keep an eye on the graphs for our next calm night.

[DaveNZ]

Calm conditions here between 5 and 6 this morning - cumulus has read the windspeed: "--- 0.0" correctly as 0.0 but has ignored the wind gusts which appear as "--- 0.0" and instead plot the last non-zero value recorded.
Hence the average line successfully hits zero but the gust line doesn't....

[beteljuice]

I thought you'd changed it so as the Davis provided the average value ?

In which case nothing has changed ?