Module 3 Week 2 During Lab

Last week you measured I-V curves for your solar panel under two different light conditions, and learned how to extract the performance metrics \(I_{\rm sc}\), \(V_{\rm oc}\), and \(P_{\rm max}\) from them. Over the next two weeks in lab, you will design and carry out a systematic investigation of how some aspect of lighting changes one or more of these metrics.

Experimental Design

Your investigation will be a success if it leads to meaningful conclusions about how lighting conditions change your solar cell’s performance. “Meaningful” conclusions pass the following tests:

• The conclusions represent something you didn’t already know from your two-data-point exploration last week.
• You can phrase the conclusions in a way that allows others to understand what they mean.
• These conclusions are a reliable foundation for predicting behavior of these solar panels in the future.

Notice that a meaningful conclusion – and thus a successful investigation – does not require you to: (a) demonstrate a dramatic (or even nonzero) effect of the quantity you are changing; (b) find an exact function that describes the way your measured quantity depends on the thing you changed; or (c) explain a theoretical reason for the dependence you found in your measurements. It also does not require you to state a hypothesis before the start of the investigation, in the sense of “we expect X to happen.” Formulating a hypothesis is one possible way – but not the only one – of making sure that your experiment is capable of distinguishing between several different situations you care about distinguishing.

To drive home the point, here are a few example conclusions that (if supported by data!) could represent successful investigations:

• Incident angle of sunlight on a solar panel has no measurable effect on the panel’s maximum output power.
• If a solar panel is shaded over its entire width but for different lengths \(y\), output power decreases slowly as \(y\) increases until \(y\) reaches half the length of the panel, when output power drops dramatically.
• As the wavelength of light illuminating the solar panel is increased, \(V_{\rm oc}\) increases steadily as a roughly linear function of the wavelength.

Just to say this one more time: the hypothetical conclusions above may or may not be true for our solar panels!

Here are a few example conclusions that would not represent successful investigations, even if supported by data:

• When we covered some of the lights with cardboard “clouds,” the solar panel got worse. Curvier clouds made the solar panel output lower, but every time we made the clouds curvier we had to cut off more cardboard from them, so the clouds were just smaller too.
• When half of the solar panel is shaded, the maximum power goes down.
• As we shine light on the panel at steeper and steeper angles, \(V_{\rm oc}\) stays constant until the illumination is 45° from perpendicular, when it increases dramatically, but it’s also possible we just missed a unit change on the voltmeter and recorded those numbers wrong.

Data Collection

During Weeks 2 and 3, you and your partner will agree on an investigation, carry it out, and analyze your results to draw a conclusion supported by the data. Be sure that you have identified one aspect of the lighting to change deliberately, while holding other potentially relevant factors constant. Explore what range of settings you can achieve for the quantity you are changing, and then design your data collection to cover that range broadly while leaving you time to zoom in more carefully on any particular settings that produce rapidly changing or otherwise important results.

Try to understand and mitigate the most significant sources of uncertainty in your determination of \(P_{\rm max}\) (or \(I_{\rm sc}\) or \(V_{\rm oc}\) if you focus on one of those), but do not get bogged down in repeated trials for their own sake. Perhaps something else is really the limiting factor in your accuracy or precision, and you should be focusing there instead! Or perhaps the changes in your solar panel’s performance are so dramatic that they are clear even with your relatively large uncertainties – and the best use of your time is to explore more different settings instead of gathering more data on just a few.

Experimental design, data collection, and data analysis and interpretation are often presented as a linear sequence of activities. In practice, they are much more cyclic, with preliminary data collection and interpretation giving feedback on the viability or usefulness of the original plan. Especially in an investigation that only spans a few hours of lab time overall, reviewing preliminary results to adjust your next steps is almost certainly more useful than adhering to a strict standard of results-blind data-taking and separate analysis.