Weather station recommendations?

[Timothy Buchanan]

I have a spare Raspberry Pi3 to run WeeWx, but no weather station yet, and would like recommendations. Reliability and accuracy are top priorities, and I'd prefer one of the models marked as tested. I am considering a Davis Vantage Vue, but would I need to buy their pricey USB interface, or could it be interfaced another way? What other stations should I consider? Thanks for all suggestions.

[Andrew Milner]

I use the vantage pro with a belfryboy logger that appears to work just fine

https://belfryboyweatherbits.blogspot.com/p/the-belfryboy-clone-usb-logger.html

[Xant]

Davis is known as the "cadillac" (prestige) of PWSs. Measurement seems precise at short time rate update. I can not advocate, as I never own one, due to pricing, outdated display and extra interface.

I went through few "budget" PWSs, that is Acurite (5x1, Atlas) and Ambient Weather (ws-1550-ip).

1) Acurite 5x1 served me for many years, till fail due to age and weather exposure --> recycle pile

2) Acurite Atlas, the natural update, had a sensor malfunction - and no extra sensors (as solar radiation) for nowadays --> returned

3) Ambient Weather WS-1550-IP its an interesting option for the price. But ObserverIP gets "overload" after just a day or two, and not suitable for WeeWX application --> returned

ObserverIP firmware problem is a known issue (even by the developer), said to be resolved, but NOT. Upon return, this was NOT acknowledged and charged for the return besides proof of difficulties (avoid!).

My current recommendation is WeatherFlow Smart PWS. At a "budget" and on top of the "usual" measuring sensors, WeatherFlow has also sensors for Radiation, UV, Illuminance and Lightning. And most important.... at NO moving parts! (similar options would cost a lot more).

https://weatherflow.com/smart-home-weather-stations/

No matter what you choose, and also to note, verify that your PWS of choice has a working WeeWX driver (very important!). All the above works with WeeWX.

Xant

[Pila]

I was recently pondering the same question and came to conclusion that price-quality seems to support my current selection. So, I am offering my facts.

Fine Offset WS1080 (sold under many names). It seems to have a problem with USB port lockups. It can be remedied by automatically restarting it. It is said this can happen on a montly basis. OK, so that is sort of half-negative. If one has the right USB Hub, WeeWX can restart Weather station should a problem arise. For now, I can not comment on USB recovery, only that the problem existed on my original station.

Shortest WeeWx read interval is 5 minutes. External sensors update the display every 6 seconds (or something like that) so it captures wind gusts. Minus: no extra sensors can be added.

Our was bought over 8 years ago. External sensors were under influence of harsh small island sea atmosphere, strong winds (100 km/h) and strong sun. And our cats who came to be fatal to it. Windex ball bearing rusted after few years - after replacing it, all was well. Display fell down almost 2 meters to a hard wood floor (a curious kitten was to blame) and got a tiny spill damage on the screen. Still works perfectly. After some years, external humidity sensor started showing 20% more, which is common for this type of sensor. Internal humidity sensor works perfectly. One external cable got damaged by the sun, can be easily fixed. Solar charged batteries (I used quality NiMh) powering external sensors, lost power over night for the first time a day before removal! Measurements are verified to be correct. Sun shield for the external temperature sensor is far from perfect but can be separated over 1 meter from the rain and wind sensors which is a huge plus! After 8 years: apart from these problems, it stil works perfectly fine.

It was replaced 3 weeks ago. Why? Cats lost the windex arrow and somehow chew off one wind speed paddle (wind speed still works perfectly fine). One cable needs to be refited into a plug. Humidity is off 20%. We replaced it with the same model, but from a different "brand". Regular EU price available to us is 130€. The new one was bought for 100€.

The old one will be reused at another location, where windex and external humidity are irrelevant. Intereseting fact: the new model from a different vendor would not recognize sensors from the old one. Apart from that, they look identicall.

I would buy it again.

[vince]

On Wednesday, October 9, 2019 at 4:55:38 PM UTC-7, Timothy Buchanan wrote:

I have a spare Raspberry Pi3 to run WeeWx, but no weather station yet, and would like recommendations. Reliability and accuracy are top priorities, and I'd prefer one of the models marked as tested. I am considering a Davis Vantage Vue, but would I need to buy their pricey USB interface, or could it be interfaced another way? What other stations should I consider? Thanks for all suggestions.

You'll have to tell us your budget before we can help significantly.

Contrary to other reports below, the WeatherFlow is 'not' accurate if you ever get rain.  That's the one big downside.  I previously posted a long summary of the pros/cons of that one on another thread the same commenter cross-posted their reply to today.

I have both a VP2 and a WF and compare the two regularly.  Davis has a long history of reliability and good customer service.  WF is new, but their 'people' are great.  It's unclear whether their reliability will be as good as Davis, but the WF is nicely constructed hardware.

Personally, I'd go Davis Vue and find a way to get their datalogger (yes, it's absurdly expensive) if you can afford it, if accuracy is your top priority.  The WF rain reports are bad to the point of just don't use any of their data.   (admittedly, they're working it, but I wonder if they're giving up on the haptic rain sensor technology).

[Joe]

I would never buy anything but a Davis2 Pro.

[Xant]

Vince

Apologies as posting was take wrongly, as there is parallel discussions on the same topic.

Pila et al

WeatherFlow is "cat" proof! No wires and no moving parts.. very sturdy construction. Also provides number of sensors that is unusual at this price range (without add-on): Temperature, Humidity, Barometric Pressure, Wind speed/direction, Rainfall, UV, Solar Radiation, Luminance, Lightning count/distance.

Yes, there might be debate regarding their rain sensor. Might be a major issue to some, but I'm not concerned.

Underneath is a WeeWX website that compares Davis Instrument vs WF (not aware whom owns; Vince, is this yours?):

http://ae3.homelinux.net:845/weewx/compare.html

It can be noted that measurements are very comparable. Specially considering that Davis (with oudated display and communication) would cost 3-4 times WF.

Xant

[vince]

On Thursday, October 10, 2019 at 10:49:23 AM UTC-7, Xant wrote:

Underneath is a WeeWX website that compares Davis Instrument vs WF (not aware whom owns; Vince, is this yours?):

http://ae3.homelinux.net:845/weewx/compare.html

No, that's the author of the (nice) weewx driver for the WF station.

It can be noted that measurements are very comparable. Specially considering that Davis (with oudated display and communication) would cost 3-4 times WF.

Again, maybe where you live, not here near Seattle.   They're within $ 75 or so in price, which to me is money well spent.  If you built your own WF console using a pi then an end-to-end WF is actually 'more' expensive than a Vue + datalogger combination here.

But yes, the Davis console is pretty bad.   I glance at it very rarely.  That's why we use weewx, to have nice web-enabled displays :-)

[Xant]

1) Davis Vue x WF system cost

Acknowledged Davis Vue to be price compatible (instead of Davis higher-end models). Also acknowledge rain to be a major issue if around Seattle (not much here in UpStateNY).

2) WF SolarRad, UV and Illuminance

I actually start questioning WF regarding SolarRad, UV and Illuminance as the plots for all 3 variables repeats every "peaks & valleys". That is, all 3 plots VERY similar from one-another, and just different by a "multiplier".

Thus, the following:

1) Are SolarRad, UV index and Illuminance different entities (which it should), or correlated by a "multiplier" (which should not)?

2) Are there actually 3 sensors in WF to measure SolarRad, UV and Illuminance separately?

Bottom line, are WF SolarRad, UV and Illuminance real measurements or "fake news"?

Note: tried many times to include a snapshot of plots comparison, but GoogleGroups automatically deleting if including image (unfortunately).

Xant

[vince]

There have been 'many' discussions about your question on the various optical sensors and reported measurements on the WeatherFlow forums.  Far too many nuances there to reiterate here.   They answered all the questions you asked above.

All the WF sensor data is basically the same as you'd expect.  There are sensors that put out a signal.  They convert it to a value.  Sometimes there's some math involved based on how your gear is tuned, or located lat/lon, etc.

There has been great 'religious view' type discussions there about sensors vs. reported values and how the reported values are derived.   I try not to get into those threads, as I trust the vendor to know what they're doing.  They kinda do this stuff large-scale for big companies and governments etc. on their many thousand dollar gear.

I did reverse engineer one pair of measurements, I forget which, and found that it seemed the default was a simple multiplier.  A few hours of research pointed me back at the reference documentation from NWS or the like that they derived their formula from.  But they 'do' salt to taste via their CL algorithms, so you can't assume that an initial (untuned) multiplier is what they're using after your station has been up long enough to get tuned by them (which they do not reveal to us).

Regardless, search the WF forums.  The specific things you've asked have been asked+answered there with lots of detail.

[Timothy Buchanan]

After reading these many good opinions and checking provided links and other leads, I've decided to buy the WeatherFlow. It was between that and the Vue. The main reported downside to the WF (rain accuracy) doesn't apply so much here in Colorado where it tends to rain a lot or not. (Today is about 4-5 inches of snow.) The main downside to the Vue seems to be the interface. I have often found serial/USB connections to be trouble and would rather work with UDP packets. Thanks again for the advice.

[Xant]

Thank you Vince to point-out Solar Rad, UV and Illumination discussions that are already posted in WF Forums.

Wow... isn't this a "heated" debate? (no pun intended)

Here is one related posting from WF Forums,"Solar Radiation, Illumance, and UV Index, is there any difference?"

https://community.weatherflow.com/t/solar-radiation-illuminance-and-uv-index-is-there-any-difference/3004

Excerpt note from WF developer:

There are actually two sensors, each tuned to a different part of the light spectrum, in the part we use to measure sunlight (the APDS-9200 from Broadcom). The APDS-9200 reports “lux” (aka “brightness”) and “uv index” directly (which must be calibrated), and we estimate solar radiation from the brightness value assuming it’s light from the sun (ie., we know the makeup of the full spectrum, more of less, even though we’re only measuring radiation from a small portion of the spectrum).

and Sensor reference:

https://github.com/CalvEngIO/APDS-9200/blob/master/Docs/APDS-9200.pdf

Long discussion and somehow confusing. Besides stating that WF has in fact 2 Sensors for Radiation, UV and Illumination, the discussion of pure "multiplier" still persists, and does not (clearly) answer the original question if there are any difference among the 3 entities.

Still... "food for thought".

Xant

[Greg Troxel]

Xant <armando...@gmail.com> writes:

> Excerpt note from WF developer:
>
> There are actually two sensors, each tuned to a different part of the light
> spectrum, in the part we use to measure sunlight (the APDS-9200 from
> Broadcom). The APDS-9200 reports “lux” (aka “brightness”) and “uv index”
> directly (which must be calibrated), and we estimate solar radiation from
> the brightness value assuming it’s light from the sun (ie., we know the
> makeup of the full spectrum, more of less, even though we’re only measuring
> radiation from a small portion of the spectrum).

That sounds pretty clear; illuminance and uv index are measured and
solar radiation is estimated.

> and Sensor reference:
>
> https://github.com/CalvEngIO/APDS-9200/blob/master/Docs/APDS-9200.pdf
>
> Long discussion and somehow confusing. Besides stating that WF has in fact
> 2 Sensors for Radiation, UV and Illumination, the discussion of pure
> "multiplier" still persists, and does not (clearly) answer the original
> question if there are any difference among the 3 entities.

The 3 things really are different, and you can't reliably and precisely
obtain any one of them from measurements of the others.

However, their point that one can estimate irradiance from illuminance
by assuming the light source sunlight (clouds etc.) makes sense, and in
particular if one had a data set containing values from their
illuminance sensor and an irradiance sensor, one could just build a
table of interpolation points. One might even be convinced that the
errors are small relative to the claimed accuracy.

For what it's worth, the Davis VP2+ has sensors for irradiance ("solar
radiation") and UV index, and Davis does not claim to report
illuminance. Points to Davis for being straightforward.

[Greg Troxel]

I should add (and I'll try to post each of my notes only once --
something is wrong where there are multiple copies of many of your
messages, so it would be good if you looked into and fixed that):

I make plots of solar radiation and UV from my archive, and decided to
plot them together and made up factor for plotting -- without claiming
that it's meaningful, just to get things at a similar scale to look at.
Basically when I started plotting, I picked a pure sunny day and
divided, and used that.

It's clear that the actual ratio of irradiance to UV index changes
between winter and summer (for the daily peaks). This isn't really
surprising, as the peak sun angle is much lower in winter and either of
these being angle dependent isn't odd, but it was interesting to see.

[Xant]

Apologies once again if you receiving multiple copies of messages, as my posting was been immediately deleted by GoogleGroups. Turned out, it seems due to image plots trying to be inserted (which I previously done at no problem). Now avoiding pics into messages. Back to main issue...

Agreed to all above Greg. Thanks for thoughts, as your EXCELLENT response well summary all conflicts and debates regarding the Radiation/UV/Illuminance issue.

[Thomas Keffer]

Your posts were not deleted. They were sent to the "moderator's folder," for review. I don't know why. In any case, I released them all this morning.

On Fri, Oct 11, 2019 at 6:16 AM Xant <armando...@gmail.com> wrote:

Apologies once again if you receiving multiple copies of messages, as my posting was been immediately deleted by GoogleGroups. Turned out, it seems due to image plots trying to be inserted (which I previously done at no problem). Now avoiding pics into messages. Back to main issue...

Agreed to all above Greg. Thanks for thoughts, as your EXCELLENT response well summary all conflicts and debates regarding the Radiation/UV/Illuminance issue.

--
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 on the web visit https://groups.google.com/d/msgid/weewx-user/19988da9-6659-49b6-88d4-3e4d936f64f3%40googlegroups.com.

[Xant]

Let me point some further features and integrations of WeatherFlow:

1) WF API

https://weatherflow.github.io/SmartWeather/api/

2) WF and IFTTT

https://ifttt.com/weatherflow

It can upload data to Google Sheets (very interesting)

Xant

[Sean Jahnig]

Thanks for the info.

I'm very interested in that weatherflow station. Does it communicate seamlessly with windguru for live wind?

Thanks

--

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 on the web visit https://groups.google.com/d/msgid/weewx-user/1c045881-5f75-48de-8fdc-d1ba9a802b36%40googlegroups.com.

[Xant]

Not much familiar (yet) with Windguru, but it seems simple to feed through WeeWX as the following:

https://github.com/claudobahn/weewx-windguru

That said, not necessary a PWS model/hardware, but upload software (https://stations.windguru.cz/)

Xant

[Xant]

After the above reference regarding Windguru, I implemented the plugin service into WeeWX and start uploading to Windguru website (easy process, and uploading in minutes).

Upon uploading PWS data to Windguru, said that account would be upgraded to PRO (as incentive for contribution and removal of Ads). It didn't happened "automatically" and I sent email to Vaclav Hornik (Czech Republic), whom seen to do it all - Windguru creator, webmaster and winsurfer (as said himself).

[Sean Jahnig]

Thanks for the update.

I contacted weatherflow regarding their station and they said they're not selling it outside of the us and Canada because of frequencies etc. That are designed only to work in those places. I guess that means it won't have full functionality outside of us and canada

Thanks,

Sean Jahnig

m.  +97150 475 4234
e.  seano...@gmail.com

On Wed, 16 Oct 2019, 19:51 Xant, <armando...@gmail.com> wrote:

After the above reference regarding Windguru, I implemented the plugin service into WeeWX and start uploading to Windguru website (easy process, and uploading in minutes).

Upon uploading PWS data to Windguru, said that account would be upgraded to PRO (as incentive for contribution and removal of Ads). It didn't happened "automatically" and I sent email to Vaclav Hornik (Czech Republic), whom seen to do it all - Windguru creator, webmaster and winsurfer (as said himself).

--
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 on the web visit https://groups.google.com/d/msgid/weewx-user/46dbbba4-cf78-4966-9d00-d83d4093e645%40googlegroups.com.

[Xant]

Not sure where is your weather pole location, but here is a bit of debate to consider.

I previously took an Acurite 5x1 to family in Brazil. Upon installing and connecting, the Acurite website was not registering "Brazil" as a location, so we had to register as US based. The result is that uploaded data was all correct (per measurements), but the forecast was to be ignored as it not really reflects the right location.

That shows dependency of your PWS hardware OEM againt their Website and controls.

But for us "WeeWX enthusiasts" this debate is futile, as all PWS data are under our control.

[Greg Troxel]

Sean Jahnig <seano...@gmail.com> writes:

> I contacted weatherflow regarding their station and they said they're not
> selling it outside of the us and Canada because of frequencies etc. That
> are designed only to work in those places. I guess that means it won't have
> full functionality outside of us and canada

I don't think that's the right interpretation.

There is a notion of unlicensed radiofrequency usage, typically for
short-range communications, and many countries have regulations about
this. However, a lot of them are different and in particular the
frequencies that are allowed are different. My understanding is that US
and Canadian rules are very similar.

In the US this goes by the name "Part 15" (from 47 CFR 15), and you will
find hobbyist parts that operate on 915 MHz here but in Europe are
supposed to be operated on 868 MHz.

https://en.wikipedia.org/wiki/ISM_band
https://www.adafruit.com/product/3070


WiFi has similar issues, with different channels in different countries.

I really don't know the details of WF, but if they say they're only
selling it in US/CAN because of "frequencies", that's a huge clue that
the device transmits on frequencies that are not necessarily authorized
in other countries.

[Xant]

Greg

Your interpretation of Radio Frequencies seems correct. Although, my note regarding OEM data upload still to consider as a personal experience in other sides of the Globe.

By other posting, it seems that Sean Jahnig already decided and bought an Acurite... which I also think (not sure if frequency to other markets), only provides to 915 MHz.

[vince]

On Wednesday, October 16, 2019 at 9:45:15 AM UTC-7, Sean Jahnig wrote:

I contacted weatherflow regarding their station and they said they're not selling it outside of the us and Canada because of frequencies etc. That are designed only to work in those places. I guess that means it won't have full functionality outside of us and Canada 

 

 

That's not quite completely accurate.   They initially did a run of over 3000 units for their kickstarter supporters, split between US and European frequency compatible gear.   After that they did additional hardware runs for the US market only, if I recall correctly.

Their 'rain check' functionality was initially rolled out US only as well, so it's possible that they simply delayed non-US additional units until they could support everybody just as well.

They also pre-announced a coming revision of the station on their website just the other day, which from the picture looks like it's all going to be in one unit seemingly built on the same kind of technology.  Of course with (another) kickstarter it's very hard to predict what the actual hardware will look like or how good it will be, especially initially.  But that's the risk of saving some money and going the kickstarter initial supporter route rather than waiting for the generally available gear to become available.

 

[Xant]

Ok... don't want to make this a tug-of-war, but my "un-scientific" fell is that you can bring any PWS hardware to anywhere, and expectations that they would communicate just fine (regardless of local regulations).

Then, there is the upload to usual OEM Weather website, which by my previous posting, some don't make much easy (which they should.... still puzzles me why they would care to restrict). But you are free with WeeWX.

I had a previous run with Ambient Weather WS-1550-IP, which it seems actually a Chinese brand, and sold under diverse brand names. The system itself is much affordable, with extra sensors, but much trouble with ObserverIP "overload". If frequency is of concern, I may assume that some version may exist.

Also assuming that most US brands to refers to 915 MHz, regardless if sold in different markets.

But honestly... is frequency a concern regardless of location?

Again, this is "un-scientific", so I can take some educated criticism...

[Xant]

Greg

Back to the Solar Radiation, UV and Illuminance.

There are indeed debate regarding direct conversion from Radiation to Illumination (as definitely different entities). Although, extrapolation by a multiplier seems to provide close results (Lux ~ 119.*Rad), per related following references:

• US Dept of Commerce / National Bureau of Standards, Solar Radiation and Illumination [1981]: https://www.govinfo.gov/content/pkg/GOVPUB-C13-42a9d8afbf69d850c07ab73935aef0ab/pdf/GOVPUB-C13-42a9d8afbf69d850c07ab73935aef0ab.pdf
• Watts to Luxes Ratio for Sollar Radiation by TMY Data for Odessa [2017]: https://www.researchgate.net/publication/318260849_Watts_to_Luxes_Ratio_for_Sollar_Radiation_by_TMY_Data_for_Odessa
  

Per your previous posting and statement: "One might even be convinced that the errors are small relative to the claimed accuracy", as building a table of interpolation points.

The above references indeed certify your correct comments... therefore, much credits to you.

Xant

[Greg Troxel]

Xant <armando...@gmail.com> writes:

> But honestly... is frequency a concern regardless of location?

Some people wish to comply with RF spectrum usage regulations.

It would not surprise me if a company attempted to avoid facilitating
violating the rules, such as by not having the web service work for a
unit that is operated in a place where it is not approved. From the US
legal POV this is perfectly reasonable behavior.