After rain yesterday my wind direction has been SE with an average speed of 5 mph with a high gust of 19 mph . My Davis has a windspeed of .3 mph, max gust of 9 mph from the WSW. Just looking at the leaves on trees the Davis is definitely right. I’m guessing this is a water problem.
Tom
1 Like
Note that I presumably have the rain on plate issue as well.
It causes the wind direction to flip radically back and forth for hours and creates bizarre wind gusts in the 40 mph range when little wind is actually present. The problem goes away after a few hours of dry weather. It has happen several times now. I haven’t attempted any work around such as protective coatings as the sensor isn’t easy to get to on my roof.
Here is a screen shot of it in action. Note that about 1/2" of rain fell between 3 and 5 am which was more than enough to cause the erratic wind direction readings and excessive wind speeds for the entire morning.(Instantaneous wind direction readings show a 180 degree change every few seconds.) Note that the actual wind direction was out of the W at about 8 mph with few gusts at no more than 20 that slowly veered to the NW. Readings returned to normal a couple of hours after the big final imaginary gust of wind occurred.
Now that mine has dried out the wind shifted to the right direction. You can see the wind shifting SE to W gradually as it dried.
1 Like
Just to complete my report, I’ve reported the issue here:
The erroneous wind speed was not really high (e.g. 7 km/h) but the air was definitely calm from end of rain through the night. One can tell the wind reading was wrong by observing gust-avg-lull. At certain points in time, they all matched. My other 3 stations did not report any wind. Even if I take into account the wind-cups inertia.
1 Like
So just as the updated title says, it looks like Frost got on the ‘plate’ early this morning of my Sky and it was doing the same thing rain was. No wind was blowing at all, but my graph has been all over the place. I just don’t know if there will ever be a solution to combat this and it disappoints me a little to think that there will always be the potential with rain or other things to give false information. I know no weather station is perfect, but I just wonder if traditional ones have less skewed results?
There will! We are actively working on a solution to this issue from both the hardware (hydrophilic and superhydrophobic nano-coatings) and software/data side!
5 Likes
This morning I took a look at what the wind was doing after last night’s storm and this is what I saw:
Stations 5080 and 5075, respectively.
I knew something wasn’t right so I went up and took a look and was surprised at how much water was on the upper surface.
Here is 5080:
Here is 5075:
I’m assuming that the water drop on the sensor on the right of 5080 is why it was reading a relatively consistent wind direction. After drying them off at about 9:55 AM on 10/28/2018 things returned to normal. Here are the 3 hour wind direction plots showing the obvious change from drying them off.
BTW, @btst.online, I love the polar wind plots.
@dsj, after seeing how much water was on the upper plates I was wondering if the top of the sky were made maybe 2cm larger in diameter if that would reduce the amount of water splash. I’m not sure how that would affect the wind flow, however.
5 Likes
I had the same idea some time ago, but with a slightly other ‘solution’. perhaps for a next generation.
In this picture the wind goes in from the side and up in the bit hollow top piece, through the sensor area and out on the other side. Wind speed would need calibration, but at least no rain in the sensor area.But I think that WF might have gone through several designs and the current one is perhaps better for other reasons
3 Likes
@gizmoev
Thanks for the shoutout. And I am very happy the PolarWind graph is working well for you.
—Sam
1 Like
That’s very likely.
An interesting idea to be sure!
This morning the wind picked up and it was starting to rain.
Wind was constantly south. But since a couple of hours the wind is all over the place, after some windgusts.
Think it’s due rain on the sensors?
Edit:
Sensors are dry now. Wind did go back to South. 
So it surely was the rain that did it.
3 Likes
perhaps construct a little hood for the sky with a hole in the center so the rainsensor still listens to the rain, but there is no possibility for the rain to blow from the side in between the sonar area. If you manage not to attach it to the sky unit itself, you should be fine, if you do attach it, it might ‘hear’ a lot more rain. If that is a problem I don’t know as the autocallibration (once deployed) might just correct for that.
I appear to have a different rain/wind related issue. During a burst of moderately heavy rain this morning, I was measuring wind speeds of 100 km/h gusting to 140 when all local stations were measuring about 1/10th of this! Furthermore, the wind direction was out by about 180 degrees and that direction error has persisted although measured speeds have returned to normal.
Rain measurement is reasonably accurate.
Do you have access or at least a view to your sky ?
Can you recognize wether there are some water drops in the wind measuring slit, either on the bottom or on the ceiling ?
I moved AIR to a higher location recently and can no longer view it, as you can see! Maybe if I stood in the pouring rain with a pair of binoculars…
I noted that a similar thing happened, at the same time, to another Weatherflow station in Southend-upon-Sea, that being the closest one to me. However, it stil retains its Wunderground gold star, which eludes me:-)
Guess you’re meaning Sky 
1 Like
Yes; sorry! Brain malfunction.
1 Like
Thanks for the report! This looks more like the “classic” rain-on-plate issue, so I’ve moved your post to the existing discussion on that topic. A quick look at your data seems to indicate one of the filters is failing. We’ll take a deeper look and hopefully your data will help us find and fix that bug.
2 Likes
That’s very reassuring; I will follow with interest.