Need for control set-up or not in this experiment?

Dear community,

Please let me know if this is the wrong category, or even the wrong forum. I post here since until now this project has involved working with Arduino.

I have been working on a device enabling to characterize the Max Power Point (MPP) and Fill Factor (FF) of solar cells by applying various loads to a solar cell and extracting the current-voltage curve. Now it is time for the experiment itself.

The idea for the experiment is to determine the FF of, say 10 cells under real sun irradiation, under sunny conditions. The parameter which can (strongly?) influence the results is that the conditions of real sunlight irradiation are highly unsteady: the proportion of diffuse and direct irradiation can vary significantly within minutes or seconds, influenced even by clouds invisible to the human eye.

In order to cope with this, various sets of measurements will be performed and they will then be compared with linear regression methods.

My question to the engineers/scientists here is this: in this kind of experiment, is there necessity for a control setup? I mean, I'm comparing various sets of measurements in which nothing changes except the illumination radiation (assuming constant temperature since experiments are conducted around the same time of the same day). But these all seem to be experimental setups, I wouldn't know which one to define as the control setup.

Thanks a lot for any clarifications regarding this question!

Yes, a control can be used to allow for differences in illumination. A single cell in the centre of the array can be used as a reference. If all cells are producing low power, then compare to some measure of the control cell to say if the array is shaded or they are actually performing poorly.

This isn't usually necessary as you can just take more measurements and apply some processing. For example, throw out any measurement when you think it might have been cloudy.

An irradiance sensor (a calibrated solar cell) can be used too.


A control in this case would be a cell receiving no sunlight.

For example, throw out any measurement when you think it might have been cloudy.

No. Bad science.

A calibrated known cell could be used as a reference.

For example, if you know of a cell with already defined and calibrated FF, you could use this to compare your own cells to under the same conditions.

Thank you all for the valuable inputs. I will indeed most probably use a calibrated cell as reference, for every measurement.

A control is needed to prove a null hypothesis. just need to vary an independent variable and collect quality data (calibrated/accurate and precise).