rtldavis windspeed problem

[Danny Ciarniello]

I have just set up weex with a Vantage Vue using weewx-rtldavis.  At first everything looked good but then I noticed that the WeatherLink Console that I have was displaying wind speeds but weewx was getting mostly zeros.  From the logs, most of the windspeed readings that the driver is producing are zero.  The occasional non-zero readings are mostly multiples of 1mph.  Why would this be happening and what can be done to fix the issue.

Here's the RtlDavis section of weewx.conf.

[Rtldavis]
   cmd = /etc/weewx/bin/user/rtldavis
   transceiver_frequency = US
   iss_channel = 4
   anemometer_channel = 0
   leaf_soil_channel = 0
   temp_hum_1_channel = 0
   temp_hum_2_channel = 0
   rain_bucket_type = 1
   debug_parse = 0
   debug_rain = 0
   debug_rtld = 2

   save_pct_good_per_transmitter = False
   driver = user.rtldavis

This is with weewx 5.2.  The Vantage Vue is the US metric version.
                                                                                         

[Danny Ciarniello]

Seems that I spoke too soon that it's only the wind that's problematic as I got a high rain rate of 204.8mm/hr with an amount of 51.2mm over an archive interval today.  That definitely did not happen and was not recorded by the weatherlink console.  I'm starting to wonder about the reliability of rtl-sdr for reading Davis weather station data.  I did not notice any problems with reading acurite 5in1 data using rtl_433 for a week or so before I got my Vantage Vue so I wasn't expecting this.

Are there any more reliable methods for reading Vantage Vue data transmissions?

[vince]

As far as I can tell it’s the only way to do so if you are reading the sensor RF directly.  There are a number of drivers that query the Davis servers if you have the new tablet-like console.  Look in the wiki in the drivers section for WeatherLink.

[Danny Ciarniello]

I'm aware of the WeatherLink driver but that driver only downloads archive data once very five minutes (by default).  What I'm looking for is something that gets the instantaneous values so that I can get a live stream for the Belchertown skin.  I was looking at meteostick but it looks like the weewx-rtldavis driver and the meteostick driver have the same roots so I'm wondering whether I would encounter the same problem with the meteostick.  I guess it would depend on the underlying sdr driver that the meteostick uses compared to rtldavis.

Ultimately, what I would really like is to figure out why the wind speeds that I'm getting aren't in keeping with what the WeatherLink Console is displaying.  The wind has picked up a bit at my place in the last few days and I'm now seeing more non-zero wind speeds but it's still mostly zeroes.  Wind gust is always a multiple of 1mph which is obviously not realistic.  

Am I the only one that is seeing this or are others as well?  If I'm the only one, how do I figure out why and how to fix it?  While I'm not particularly familiar with Python and not at all with golang, I am willing to do whatever it takes, with the help of the experts in this group, to figure it out.

[vince]

No ideas here. I rebuilt mine on a spare pi and will leave it running since we have a week of storms coming to the Pacific NW. I see a few duplicate packets logged and very occasional missed packets but I haven’t noticed craziness yet.

You can probably filter out wild readings with some StdQC settings if you just want to ignore them.

[Greg Troxel]

Danny Ciarniello <dan.cia...@gmail.com> writes:

> Ultimately, what I would really like is to figure out why the wind speeds
> that I'm getting aren't in keeping with what the WeatherLink Console is
> displaying.

> Am I the only one that is seeing this or are others as well? If I'm the
> only one, how do I figure out why and how to fix it? While I'm not
> particularly familiar with Python and not at all with golang, I am willing
> to do whatever it takes, with the help of the experts in this group, to
> figure it out.

weewx-rtldavis surely just interprets output from rtldavis, and I think
you should run rtldavis (written in go?) by itself, without weewx, and
log the output, and see what that looks like.

[Danny Ciarniello]

I have the same set of storms headed my way given that I'm in the BC Lower Mainland. :-)

From what I see in the logs, I thought that rtldavis output "raw" data while weewx-rtldavis interpreted it; I don't see any setting for rtldavis that would convert the output to human-readable values.

The attached log contains a one minute log excerpt from weewx where I've turned on debugging for both weewx and rtldavis.  Unfortunately, both rtldavis and weewx-rtldavis output is flagged with user.rtldavis, which makes it difficult to differentiate but I know that lines with "m1:" and "data:" are from rtldavis.

All entries in the excerpt have 0.0 for the wind speed but, going by what was displayed on the WeatherLink console, there were non-zero speeds at

20:51:11 - 1.5 kph

20:51:14 - 0.5 kph

20:51:19 - 0.4 kph

20:51:28 - 1.1 kph

20:51:30 - 1.5 kph

20:51:34 - 1.4 kph

20:51:36 - 1.1 kph

20:51:39 - 1.0 kph

20:51:41 - 0.7 kph

20:51:46 - 0.6 kph

Going by previous behaviour, my assumption is that if any speeds were above 1.6 kph (1 mph) I would have seen non-zero wind speeds.

[John Smith]

On Mon, 8 Dec 2025 at 17:49, Danny Ciarniello <dan.cia...@gmail.com> wrote:

20:51:11 - 1.5 kph

20:51:14 - 0.5 kph

Not sure about your setup etc, but my Davis weather station only reports whole km/hr values, are you sure those values aren't m/s or some unit other than km/hr? 

[Danny Ciarniello]

I have configured my console to display data with one decimal place precision and wind in kph so yes, I'm sure.

[John Smith]

Units transmitted are usually fixed and don't have to match those being displayed.

--
You received this message because you are subscribed to the Google Groups "weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to weewx-user+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/weewx-user/8ec1c9e0-fd81-4770-8ec9-0a2f11c09c75n%40googlegroups.com.

[Danny Ciarniello]

The issue is not one of units; it's an inconsistency between what the Weatherlink console is displaying and what weewx is.  

For example, for the series of speeds that I posted previously, which were taken from the console, weewx recorded 0 as the speed.  From observation, when the console displays a speed between 0 and 1.6kph, weewx records 0; for speeds between 1.6 and 3.2 kph, weewx displays 1.6kph, for speeds between 3.2 and 4.8kph, weewx displays 3.2kph and so on.  Obviously, if those values are converted to mph, the progression is 1mph, 2mph, 3mph which implies that for some reason, weewx is only recording speeds in increments of 1mph.  A complication is that while that is generally what happens, I do occasionally see weewx recording wind speeds that are not multiples of 1mph.  Wind gust, however, is always a multiple of 1mph.

The rtldavis/weewx-rtldavis and the console are obviously receiving the same signal but they are interpreting it differently.

[John Smith]

On Tue, 9 Dec 2025 at 17:44, Danny Ciarniello <dan.cia...@gmail.com> wrote:

The issue is not one of units; it's an inconsistency between what the Weatherlink console is displaying and what weewx is.  

Then what is it?

 

For example, for the series of speeds that I posted previously, which were taken from the console, weewx recorded 0 as the speed.  From observation, when the console displays a speed between 0 and 1.6kph, weewx

Which could easily be explainable as rounding somewhere...

 

records 0; for speeds between 1.6 and 3.2 kph, weewx displays 1.6kph, for speeds between 3.2 and 4.8kph,

To see 1.6, 3.2 and 4.8km/hr sounds exactly like it's converting from whole mph readings...

 

weewx displays 3.2kph and so on.  Obviously, if those values are converted to mph, the progression is 1mph, 2mph, 3mph which implies that for some reason, weewx is only recording speeds in increments of 1mph.  A complication is that while that is generally what happens, I do occasionally see weewx recording wind speeds that are not multiples of 1mph.  Wind gust, however, is always a multiple of 1mph.

As I said, mine always reports whole numbers, but in km/hr, never fractions...

 

The rtldavis/weewx-rtldavis and the console are obviously receiving the same signal but they are interpreting it differently.

Exactly what I said... What the console shows doesn't have to match the units of the values in the background... 

[vince]

I’m running two weewx instances as a test.  One uses the Davis serial datalogger attached to the old style console and the Vantage driver. The other uses a SDR to sniff the RF and the rtldavis driver.

In a couple wet/windy days testing the NOAA reports are identical for temperature and rain. The wind readings are very similar but not identical, differing by only a fraction of a mph in high speed and average speed and differing slightly in dominant direction.

I can look a little deeper with a few days more data, but I don’t see crazy differences here so far.

[Danny Ciarniello]

To make the issue more clear, I have recorded a video that shows the issue that I am trying to resolve: video

As can be seen, the wind speed displayed by the console is not reflected in that displayed by weewx   weewx is only show speeds that are multiples of 1.6 kph (i.e. 1mph).

Here is an excerpt of what's in my DB for wind speeds and wind gust.

MariaDB [weewx]> select round(windSpeed,1) as windSpeed, round(windGust,1) as windGust from archive order by dateTime desc limit 10;
+-----------+----------+
| windSpeed | windGust |
+-----------+----------+
|       0.4 |      2.0 |
|       0.3 |      2.0 |
|       0.6 |      4.0 |
|       0.3 |      2.0 |
|       0.9 |      4.0 |
|       0.7 |      3.0 |
|       0.8 |      4.0 |
|       0.6 |      4.0 |
|       0.4 |      3.0 |
|       0.4 |      3.0 |
+-----------+----------+

The wind gust is obviously in increments of 1mph while the wind speed shows more variation.  Based on what I've seen in weewx, I did not think that there were any non-zero wind speeds being recorded but I was obviously mistaken about that.  Nonetheless, those speeds are not being displayed in the live widget.

Basically, my expectation is that, within reason, the speeds displayed by weewx and that by the console are the same.

In case it's not obvious, the skin that I'm using is Belchertown with mqtt integration.

[vince]

I think you're seeing driver vs. US-vs-metric conversion issues perhaps....

• The Vantage driver via the old-school console seems to round(0) both windSpeed and windGust
• The rtldavis driver via SDR rounds windGust, but not windSpeed
• The WeatherFlowUDP driver for their Tempest station does not round either item

VP2 via serial logger and old school console - US everything...

echo "select datetime(dateTime,'unixepoch','localtime'),windSpeed,windGust
    FROM archive ORDER BY rowid DESC LIMIT 5;" | sqlite3 vp2.sdb
2025-12-09 13:55:00|0.0|0.0
2025-12-09 13:50:00|0.0|0.0
2025-12-09 13:45:00|0.0|1.0
2025-12-09 13:40:00|0.0|1.0
2025-12-09 13:35:00|0.0|0.0

VP2 via SDR...

echo "select datetime(dateTime,'unixepoch','localtime'),windSpeed,windGust
    FROM archive ORDER BY rowid DESC LIMIT 5;" | sqlite3 archive/weewx.sdb
2025-12-09 13:45:00|0.0173913043478261|1.0
2025-12-09 13:40:00|0.112068965517241|1.0
2025-12-09 13:35:00|0.0|0.0
2025-12-09 13:30:00|0.00854700854700855|1.0
2025-12-09 13:25:00|0.170940170940171|2.0

WeatherFlow Tempest from emitted 'metric' UDP converted by the driver to US units

echo "select datetime(dateTime,'unixepoch','localtime'),windSpeed,windGust
    FROM archive ORDER BY rowid DESC LIMIT 5;" | sqlite3 tempest.sdb
2025-12-09 13:40:00|0.948463345222265|1.6105981333963
2025-12-09 13:35:00|0.957411112630022|1.43164278524115
2025-12-09 13:30:00|1.02004548448432|1.96850882970659
2025-12-09 13:25:00|0.903724508183479|1.6105981333963
2025-12-09 13:20:00|0.769507997067121|1.38690394820237

[Danny Ciarniello]

I suppose it's possible that it's US/Metric conversion issues though, as far as I know, the only difference between a fully US version and mine is the rain tipping bucket (0.01in vs .2mm).

If you're right about the rtldavis driver, it looks like I need to figure out the code and learn how to build a golang project.  :-)