Your browser does not support script
Home Search Disclaimer Login/Register
Updated : 16/10/2016

GHowSAW Weather Radiation Calibration


There are numerous radiation sensors available but intially we will be dealing with the SFH203P photo-diode typically supplied by HobbyBoards. The photo-diode is operated with reverse bias to ensure a linear response to the applied input light and to improve the response time. The load resistance is 390 Ω. Under such conditions we can refer to the response data shown below to derive the transfer function.



Transfer Function

From this response curve we can determine the transfer function of the device as being:-

Isc = (7.8410-3Ev+0.3366)10-6

Isc = 7.8410-9Ev+0.336610-6

Where Isc is the short circuit current in amps and Ev is the incident energy in lux (lum/m2).

Since Isc is directed through a load resistor of 390 Ω then the resultant voltage Vo will be given by:

Vo = 390 (7.8410-9Ev+0.336610-6)

Vo = 33.057610-6Ev+1.312710-4

Ev = (Vo106-131.27) / 3.0576


The HobbyBoards radiation sensor uses the "Battery Current" feature of the DS2438 slave device to measure the current of the SFH203P by measuring the voltage generated over a 390 Ω resistor via the Vsens+ and Vsens- pins. The DS2438 current sensor accepts a maximimum voltage across the Vsens+ and Vsens- pins of 0.25V and measures this voltage via a 10-bit ADC thereby providing a resolution of 0.2441 mV (2.4 10-4 V).

The maximum measurable lux level is 81 720 lux and the resolution is 78 lux. This equates to a maximum of 875 W/m2 which may be a problem for my location in South Africa since the insolation levels can exceed 1000 W/m2 in mid summer. Once I have a full year of detail data I may have to re-assess the calibration method or adjust the value of the load resistor to extend the range.



Directional Response Correction

Because the directional response of the SFH203P is not uniform (see figure below) a correction is required based on the solar zenith angle. This correction is required to implement the algorithms for bright sunshine and cloud detection and classification

To simplify the calculations the directional charcteristics curve is divided into two regions of 0-70° and 71-90° and respresented by two equations for the directional factors.

For Zenith angle of 0-70°

y = -1.45082410-6x3 + 2.43051110-6x2 - 1.36646910-4x + 1.0

For Zenith angle of 71-90°

y = -0.0014x + 0.6

Where y is the directional factor and x is the zenith angle. For the equation for zenith refer to solar position calculations


The SFH203P may be calibrated by direct comparison with a commercial lux meter as shown at 1-Wire Solar Sensor Calibration. Whilst this site dicusses the calibration of a HobbyBoard Radiation Sensor with a Lux meter the unit tested may be an older version since the site discusses voltage outputs up to 7v which is far in excess of the 250mV range of the current sensing pins of the DS2438.

Personnally, I just check the device against theoretical values using the calculator to ensure that the device is set up and working correctly and that I get consistent results. (Note: The calculator it set up to calculate solar radiation not illuminance so divide the result in kW/m2 by 1.353 kW/m2 and multiply by the terrestrial solar illuminance constant of 127 500 lm/m2 (lux) to get a result in lux.)

With a model of the daily solar irradiation such as Bird and Hulstrom's model from the publication "A Simplified Clear Sky model for Direct and Diffuse Insolation on Horizontal Surfaces" by R.E. Bird and R.L Hulstrom, SERI Technical Report SERI/TR-642-761, Feb 1991. Solar Energy Research Institute, Golden, CO. used in conjunction with the NOAA JavaScript solar position calculator.(N.B. if using this javascript please note the reverse convention used for longitude and time zone) the expected solar irradiation for can be calculated. Using the methodology described in the page on cloud identification any period of clear-sky condition can be identified. During a period of clear-sky conditions a direct comparison of the measures and predicted irradiation level will indicate if any adjustment of the sensor calibration is necessary. The optimal time for this comparison is the solar noon. With this approach it is possible to automate the calibration process. At present the GHowSAW system carries out this comparison and logs the relevant values but does not automatically adjust the calibration. When sufficient data has been accumulated (at least daily results for over a year) the automated process may be activated if the data indicates that regular adjustments are necessary.

528 Page Hits