[SR123]

Thanks Nigel,

I am glad to share my ideas

I had my fine offset station, an Aercus WS2083, for five years and in that time I've had to change the bearing in the wind speed sensor, a reed switch in the wind direction and the reed switch in the rain collector!
I don't know if your FO station is connected to a pc of some kind but mine has been connected to various types over the years, a windows mini pc, a raspberry pi4, a pi Zero and now a Zero 2. Since moving over to WeatherDuino I have not had any lock ups of the pi which I was getting frequently with the fine offset station I was also getting quite a few 'unable to connect to web page' errors when trying to connect to the Cumulus MX dashboard - all those problems have disappeared too.

The only thing I miss from the Aercus display is the ability to just glance from across the room to see the outside temperature, now I need to put my glasses on ! Maybe something Werk can play with making the temp display bigger when on the 'Big Clock' page.

Thanks, Steve

Hi SR123,

I've really enjoyed reading about your Stn.  It's given me some ideas.  I do like what you have done with your RX been thinking about what to do with my own.  Looks like some bits of roofing batten (38 x 50 mm) might come in useful...  Though I have been thinking of something else.  I was intrigued by the strainer that you mounted your T/H sensor in.  I too have been using a FO system for a number of years and having replaced a number of parts / whole lot have decided to abandon FO.  When I inspected the TX over winter break it was badly corroded and I do what to protect the sensor more than in the FO set up.  I have retained the radiation screen and have been thinking of ways to use this with additional protection for the little PCBs.  I have some "conformal coating" to put on the PCBs (except the sensors that is).  I might emulate your strainer screen though I have been thinking about a tube inside the radiation screen...

I've gone for a bought antena both TX and RX and will keep the TX one in the IP enclosure so as to minimise holes in the enclosure (though conversely I plan to have the indicator LEDs visible in the side of the enclosure!!!).

I've not have the solar data before so am only just giving it some thought.  As Werk is saying fused silica is the best for UV transmission it is used in UV reactors for disinfection of water.  Despite being surrounded by United Utilities and Yorkshire Water reservoirs my house is on a private water supply!

Stay safe

Nigel

[AllyCat]

Hi,

... Today I decided to try different materials to cover my solar & uv sensor. ....... then 2mm clear plastic - not sure what it is, it was from a box lid,

The graph in post #14 is very interesting and confirms my suggestion in #9 that Perspex (Acrylic sheet) is not suitable as a cover/protector for (at least) the UV sensor. But we can learn much more from the graphs: First, it appears that the "2mm plastic" is NOT "Perspex" (because considerably more UV is reported); Google indicates that there are about 8 different "Clear" Plastics, of which about half are transparent also to UV. However, some of these plastics are not even "resistant" to UV, so it would be very unwise to choose a material simply because it looks transparent!

Then, note that the Ratio of w/m² against UV is different for the two "clear" (uncovered) periods, and also that it changes during the time after 14:00.  This suggests that one of the sensors (probably the UV) has a different angular response to the other. That the 3mm and 2mm glass (and the 2mm "plastic") all have the same transmission is not surprising, because they are presumably all "100% transparent", but about 8% of the light is reflected by their two outside surfaces. A problem is that the proportion of this reflection increases as the sun moves close to the horizon, probably up to about 50% (lost) at around 10 degrees above the horizon (Google: Fresnel Equations). Thus the sensor is unlikely to give a good "Global" Sine Response, which might lead to an incorrect indication of whether the Sun is "shining", when it has a low elevation above the horizon:

This becomes almost a "philosophical" question: Do we want to measure if "The Sun is Shining" within an hour or two of Sunrise/Sunset? If you are interested in "Sunbathing" or perhaps the energy output from your PV Solar Roof Panels, then perhaps not. But if you wish to use it as a (meteorological) indicator of Daily "Cloud Cover" then it needs to be recorded accurately.

Note that "Anti-Reflection Coatings" (even if available) are not very appropriate here, because their operation is dependent on the wavelength of the light. That's why camera lenses look "Pink" or "Purple", because it's mainly the Green light which is NOT reflected and Red and Violet are still reflected (to some extent). This is where PTFE comes of interest; it is NOT "transparent", but it is an excellent diffuser of light, i.e. it reflects (some) light, but independently of the angle of incidence. Thus it can be arranged to indicate a genuine Global (Sine Response) Solar light level, and perhaps even correct for a poor angular response of the sensor(s).  Of course the sensitivity (or amplifier gain) of the circuit would need to be increased (i.e. a change in the circuit design), but the PV panel in particular should easily be able to deliver sufficient current (measured across a somewhat higher value resistor).

An interesting aside: The original Fine Offset "Solar Pods" had their light sensor underneath a "white dome", but they appear to have "forgotten" this when they re-launched their new (generally improved) stations under the Ecowitt (and other) Brand names.  However, it appears that the latest Ecowitt sensors have been "improved" by adding a "white filter" above their Lux sensor !   But sadly, it's also been reported (and appears to be true) that their "UVI" measurement is simply the w/m² value divided by ~100 !
_____

IMHO the graph in post #11 is quite "confusing": It has a partially Linear/Logarithmic wavelength scale and also the middle "Yellow" band does NOT correspond to the Visible Spectrum! The "Violet" region on the left has a linear scale which corresponds basically to the UVC wavelengths which do NOT pass through the Earth's atmosphere, so are irrelevant here (the UVA and UVB wavelengths which do reach the Earth's surface are in the Yellow region). Similarly the "Near InfraRed" wavelengths from 700nm to about 1000nm are not shown in the "Pink" region. So personally, I'm not convinced that the high cost of the Fused Silica Quartz Glass is justified in this application (and the "ripple" in the JGS-1 response at 1000nm might be more significant than what happens below 270 nm ! ).

In passing, Google found This PV Spectral Response Graph , but that whole web document appears to contain a wealth of useful Educational Information (all of which I haven't actually read yet).

Cheers, Alan.

[SR123]

Hi,

Thanks for the info on software settings, but before I delve into that I decided to do some more experiments

I found this info on the UVM30A sensor
   

This morning was cloudy / overcast, I measured the voltage at the tx pcb terminals for the UV - 174 mV
This according to the above graph is less than 1 UV - the rx showed 2.8 UV

This afternoon the sun came out, I measured the voltage and got 400 mV
This should correspond to 3 UV - the rx showed 6.6 UV

To me something isn't quite right, the readings from the sensor seem to be correct. Any Ideas?

Thanks for clearing that up

Today I recorded a maximum UV of 7, which is too high. The maximum should have been 5.
What would you recommend I do?

Thanks, Steve

Hi,

According to specs, the UVM-30A tolerance is +- 1UV Index.
You can also try to do some adjustments on the software side.

On the receiver software, RX_TX tab, find this code (around line 360) :
 

Code:

#if (UV_SENSOR == 1) // --- UVM-30A
IndiceUV = ((RX_Data[2] * 12.0 * (100.0 / UV_Transmission)) / 1024.0);
#endif

Try lower the "12.0" value.

Rgs.
Werk_AG

[Nigel]

Steve,

Hum just got a goat called Steve...

Mine was the W1080 from PCE well it was kind of difficult to tell.  For lock down it didn't work most of the time and I had moved the TX close to the house disconnected from the wind / rain sensors...  I used to have it connected to a windoz XP machine in the loft (and the caravan we lived in for 3.5 years prior to that), so not to worried about glancing at the values.  This used to serve to a website that a prio ISP kindly left running for me after I left them (that has now stopped).  That old machine died.  I then started my Pi journey with a Banana Pi (not great) then quickly to a Pi Zero W (excellent).  I used to use Cumulus but don't get on with the MX version and now have moved to weewx (though I see that the Davis protocol can cause issues).  I have the little pi running: weewx, Apache to serve the pages, Mosquitto MQTT, and a sense hat...  I have a number of plans but hey I need to get the weather stn up first...  There are a load of other things in life such as broken cars, livestock fences and digging in the garden that pull on any free time.  I see a number of potentially Lancastrian sites I don't know if one of those is yours.

As to glancing over at the temp...  Surely that is a little bit of personalisation in the Arduino code?  Not having it running yet I currently believe that when not prodded with the stylus the TFT is in big clock mode...  Could that be big temp mode for you?

Stay safe

Nigel


 

Thanks Nigel,

I am glad to share my ideas

I had my fine offset station, an Aercus WS2083, for five years and in that time I've had to change the bearing in the wind speed sensor, a reed switch in the wind direction and the reed switch in the rain collector!
I don't know if your FO station is connected to a pc of some kind but mine has been connected to various types over the years, a windows mini pc, a raspberry pi4, a pi Zero and now a Zero 2. Since moving over to WeatherDuino I have not had any lock ups of the pi which I was getting frequently with the fine offset station I was also getting quite a few 'unable to connect to web page' errors when trying to connect to the Cumulus MX dashboard - all those problems have disappeared too.

The only thing I miss from the Aercus display is the ability to just glance from across the room to see the outside temperature, now I need to put my glasses on ! Maybe something Werk can play with making the temp display bigger when on the 'Big Clock' page.

Thanks, Steve

Hi SR123,

I've really enjoyed reading about your Stn.  It's given me some ideas.  I do like what you have done with your RX been thinking about what to do with my own.  Looks like some bits of roofing batten (38 x 50 mm) might come in useful...  Though I have been thinking of something else.  I was intrigued by the strainer that you mounted your T/H sensor in.  I too have been using a FO system for a number of years and having replaced a number of parts / whole lot have decided to abandon FO.  When I inspected the TX over winter break it was badly corroded and I do what to protect the sensor more than in the FO set up.  I have retained the radiation screen and have been thinking of ways to use this with additional protection for the little PCBs.  I have some "conformal coating" to put on the PCBs (except the sensors that is).  I might emulate your strainer screen though I have been thinking about a tube inside the radiation screen...

I've gone for a bought antena both TX and RX and will keep the TX one in the IP enclosure so as to minimise holes in the enclosure (though conversely I plan to have the indicator LEDs visible in the side of the enclosure!!!).

I've not have the solar data before so am only just giving it some thought.  As Werk is saying fused silica is the best for UV transmission it is used in UV reactors for disinfection of water.  Despite being surrounded by United Utilities and Yorkshire Water reservoirs my house is on a private water supply!

Stay safe

Nigel

[SR123]

Think I have sorted my UV over reading problems   I searched the forum and found a post 'UV sensor for WeatherDuino Pro2 weather station'
in which the same problem was occuring.
Development team member Hornychz came up with a different formula to calculate the UV Index. I have slightly modified it to this,

IndiceUV = (( RX_Data[2] * 1.074 / (UV_Transmission * 0.01) - 135 ) * 0.011);

and changed the UV transmission in the config options to 95%

So far so good, currently reading UV of 3.9  at 13:00 (met office predict a high of 4)

Any further tweaking I will do in Cumulus MX calibration settings,

Thanks, Steve

[werk_ag]

I remember quite clear of that formula, my friendship with Hornychz started with a discussion about that.
Since then the Hornychz perfectionism has been very important on the whole WeatherDuino project.

Indeed that formula can produce better results, specially when readings are on the lower scale of the UV Index.

Rgs.
Werk_AG