I would like to measure irradiance of sunlight [W/m2] with Arduino (or at least its approximation) but I am not sure which sensor to use.
Irradiance of sunlight can be as high as 1361 W/m2 (solar constant) but on the Earth surface it is typically ~1000 W/m2 on sunny day.
I look for TSL235R sensor:
But in the datasheet, page 2, bottom, you can see that max output frequency is 300 kHz which correspond to (page 3, Figure 1) approx. 500 uW/cm2 or 5 W/m2 (since [uW/cm2] / 100 = [W/m2]). So this sensor is not appropriate to measure irradiance as high as 1000 W/m2.
If you have some suggestions which sensor would be appropriate for this application, please suggest. I cannot afford expensive solution so also cheap sensors for some other application may be sufficient to have approximate irradiance of sunlight.
for my purpose it is OK if accuracy is +-50 W/m2 or even more. I don't have calibrated irradiance meter but I have data from meteorological measurement station a few kilometers away so I can somewhat compare those measurements. It is for outside with different temperatures. But it is meant to measure only in summer and on daily time when air temperatures are around 30 °C at noon.
TSL2561 can measure up to 40,000 lux so there is similar problem like with TSL235R sensor (direct sunlight is not measurable):
It is interesting idea to measure irradiance with solar cell. If I understand correctly: The output of the solar cell (voltage) is connected with sun irradiance. Maybe correlation is simple enough that some (i.e. linear) relationship could be established. But I need some reliable comparisons with real pyranometer. After a few measurements I can try to establish such relationship and make some additional test comparisons with pyranometer.
I would propose give the TSL235 a defined set of "sunglasses" that temper the amount of light with a known factor.
The most accurate way is to have a diverging lens and mask it with a black paper with a small hole.
The sunbeam coming thought the black hole will be diverted e.g. over an area 1000 times as big as the hole.
so the intensity is also less.
Now you can place the sensor somewhere in the light cone and calculate the real value from the measured value.
LOng ago That I did math with lenses and I recall there will be a bit of gonio (sin cos) in it
However if it is to measure the radiation of solar cells in a fixed position it is preferred to have the sensor in the same orientation.
Then the angle of radiation is identical.
Thank you for your interesting suggestion: black hole with lens. I find it difficult to get proper lens and to calculate necessary.
Regarding solar cell: can I connect it directly to the Arduino analog input pin? I.e. one voltage wire from cell to A0 and other wire to Arduino ground? Nonlinear response can be dealt with nonlinear equation.
I also consider two other solutions: ~5x5 cm metal thick plate, directed into sun, well insulated below, with temperature sensor (i.e. DS18B20 connected to Arduino) below the plate. Temperature of the plate is connected with solar irradiation and temperature of the air and procedure to determine such connection is similar to the before mentioned procedure for solar cell.
Maybe also welding glass before TSL235 would do the trick and, again, correlation with known measurements must be established.
The solar cell can be connected directly to the Arduino, but I would use a series resistor of 1k. Just for safety, in case you connect something wrong.
You could set the analogReference( INTERNAL) to 1.1V. Wait 20ms after setting that to let the voltage reference become stable.