[strada916]

Hey Guys
I have read most of these posts and was wondering if there was something already available and at low cost. Looks like SparkFun have something we could consider instead of re-inventing the wheel
Here is the link to the UV sensor and there are also other sensors for Luminosit Click on the link >> UV Sensor

My 2 cents worth

Nick

[uncle_bob]

Here is the link to the UV sensor

Hi Nick,
Getting UV readings is easy already. Getting Solar Radiation readings is the tricky part
Take a look here:
http://www.meteocercal.info/forum/Thread...r-Stations

[AllyCat]

Hi,

do i reduce the gain if the data always hitting max limit, ie 1023 very frequently?

Yes.

Getting UV readings is easy already. Getting Solar Radiation readings is the tricky part

Yes indeed. But I'm not convinced that even the GUVA-S12SD (alone) is strictly the "correct" sensor either, for the "official" Erythemal UV index. Although described as a "UVB sensor", it's peak response appears to be well into the UVA part of the spectrum (and one of it's applications is described as for "UVA lamp monitoring").

The "official" UV Index almost totally excludes UVA rays, requiring an expensive filter or a specialised sensor. With the exception of Davis (who appear to use such a filter), it seems that most "UV Index" sensors (and apparently even some quite "learned" articles and reports) assume a correlation between the required measurement around 300 nm (UVB) and an actual measurement peak around 350 nm (UVA).

Of course, such a correlation may be valid in some circumstances, but the Earth's atmosphere is highly variable around the 300 nm region. It's rather like "measuring" the weight of potatoes grown in your garden in a particular year by counting the number of apples on the tree.

Cheers, Alan.

[werk_ag]

Hi,

Acording to Solar erythemal ultraviolet radiation, the UV spectrum is divided into UV-C (100 - 280 nm), UV-B (280 - 320 nm), and UV-A (320 - 400 nm) radiation.
Considering that the wavelengths responsible for carcinogenesis are mostly bellow 300nm, your considerations are important.
Anyway, I belive that the readings obtained with a GUVA-S12SD sensor, are not very different than the ones obtained with a Davis.
It would be interesting see a side by side graph of each one for the same place.
I don't have a Davis weather station, but there are some nearby me, and using wunderground, I frequently check their readings against the ones I have, and I never see a great discrepancy. Of course I only do these comparison in days with a very clear sky, without any clouds.
I think that, per example, a reading of 7.4 and 7.8 which represents a 5% difference, may be an important difference for a scientific instrument, but is not for an amateur weather station.

Yesterday my weather station located near Lisbon - Portugal, has registered a maximum UV index of 8.1, not too bad if compared to image below.



Maybe we should open a new thread to talk about UV sensors.

[seth22]

hi,

after taking my setup few days under the sun, some of the the cf only produces 0 , 5xx and 1023 values (like not sensitive at all), i'd check for all connections, none of them have loose ends or detached from the circuit, what could be the reason behind this?

[AllyCat]

Hi,

@Seth: You will have to test all the PV panels and the amplifiers, etc. using a multimeter. Measure the voltage acrross the PV panel (and its load resistor), both when disconnected and when connected to the amplifier (the voltage should be he same). With the panel connected, check that the output from the amplifier (and input to the ADC) is about 7 times the input voltage.

If suitable output from the PV panel is not available, then you may need to test with a potentiometer on the amplifier and/or ADC inputs. One end of the pot to ground, the other through a suitable resistor to a supply rail (or battery, etc.).
______

@Werk: Yes, I expect that much of the time the nearer UVB (300 nm) and UVA (350 nm) measurements will correlate quite well. But it's the "unusual" conditions that might be of interest. What about light haze (or perhaps dust from an Icelandic volcano), or under the "hole in the Ozone Layer" in Australia (e.g. does a normally calibrated GUVA-S12SD in Australia ever indicate 17 ? ). My suspicions about the FO Solar sensor were aroused when mine indicated 11 behind glass! Not only is 11 ridiculously high for the UK, but UVB rays are known not to pass through glass!

Cheers, Alan.

[werk_ag]

@Werk: Yes, I expect that much of the time the nearer UVB (300 nm) and UVA (350 nm) measurements will correlate quite well. But it's the "unusual" conditions that might be of interest. What about light haze (or perhaps dust from an Icelandic volcano), or under the "hole in the Ozone Layer" in Australia (e.g. does a normally calibrated GUVA-S12SD in Australia ever indicate 17 ? ). My suspicions about the FO Solar sensor were aroused when mine indicated 11 behind glass! Not only is 11 ridiculously high for the UK, but UVB rays are known not to pass through glass!

Cheers, Alan.

Hi, Alan

For that special or "unusual" conditions, most of the common sensors or maybe most of the amateur weather stations are also useless, including Davis, I think.

My intention with this sensors is only to allow an hobbyist like me, in a simple and not expensive way, to have solar rad. and UV data, with some tolerable degree of error, but not as ridiculous as reading a UV index of 11 in UK (unless it was one one of those special occasions )

Regards, Werk_AG

[AllyCat]

Hi,

....also useless, including Davis, I think.

No I don't think so. Davis may use a similar sensor, but they do add the correct "Erythema Action Spectrum" filter, which perhaps explains why their sensor is so expensive (hundreds of Dollars/Euros/Pounds). See the Data Sheet here, noting the spectral response at 360 nm, and the logarithmic vertical scale.

Perhaps the point I'm really making is: "When (and how) does the 'UV Index' (measured at 360 nm) deviate from a number obtained by simply dividing the 'Visible Watts/m2' value by a suitable factor (in the range 100 - 120) ?". I genuinely don't know and am interested in the answer.

However, to be more positive, my own experimental "Solar Sensor" uses two channels, each with a BPW34 sensor, one "normal" (Visible + IR) and the other has an integral "IR Filter" (which actually means it is sensitive only to IR). Comparison of the two channels has been very "interesting" to me, but is perhaps no more relevant to general "Weather Station" measurements than is (IMHO) a UV measurement at 360 nm.

But to really get back "on topic"; I have been looking at using a PTFE (Teflon) "window", to protect the sensors (particularly for BPW34 and similar devices) and also to achieve/retain a good spatial "cosine response". However, all the published data that I found referred to the "mechanical" and physical properties of PTFE, not its optical properties. In particular, I was concerned (from various web comments and my measurements on the extremely thin "PTFE tape" used to seal water/gas pipe connections) that it might be too opaque.

So eventually, I purchased a small piece of 1 mm thick "M300 PTFE sheet" on ebay at a modest cost and can report that this sheet has an optical transmission factor (transparency) of almost exactly 50%. Thus it seems highly suitable for use as a "window" to protect any solar sensors against the elements, and to achieve a good Cosine response from the sensor(s). But I have yet to accurately measure the spatial response of my own sensor.

Cheers, Alan.

[werk_ag]

Hi,

....also useless, including Davis, I think.

No I don't think so. Davis may use a similar sensor, but they do add the correct "Erythema Action Spectrum" filter, which perhaps explains why their sensor is so expensive (hundreds of Dollars/Euros/Pounds). See the Data Sheet here, noting the spectral response at 360 nm, and the logarithmic vertical scale.

I should have answered early, but ...
Thank you for this clarification.

However, to be more positive, my own experimental "Solar Sensor" ...

It seems you do things in a very scientific way. Maybe sometime you share with us your experiments... I would like, and certainly would learn a lot.

[seth22]

in post #2, why it is mentioned that the Rsh should be as small as possible so that the current is linear to the amount of sunlight strike on the cf, but the calculations of Rsh is greater than 1.5R? does the current produced still behave proportionally to the incident light?