Getting more power from your solar panels

Results

Figures 7 and 8 show the results from the March 31 and April 1 runs. Intermittent clouds on March 31 (Figure 7) created the up and down pattern of the data. After doing the math on the power generated by the two panels on March 31, I calculated that the tracked solar panel generated 21.88W-hr and the untracked panel generated 17.52W-hr, which is a 24.9 percent increase in power generated by tracking, or pretty close to the estimate of 30 percent. The test on April 1 (Figure 8) produced a similar increase of 23.8 percent.

Figure 7: March 31 power curves.

Figure 8: April 1 power curves.

Looking at Figure 8, you see that the tracked panel peaked at about 11:30, whereas the untracked panel peaked at about 12:45 (approximately local noon, when the sun is highest in the sky). Therefore, the tracking software is off about 16 degrees, which can be easily corrected to yield even better numbers. Just eyeballing the graph in Figure 8, I can see getting another few percent improvement, so it would come even closer to the 30 percent expected increase in power generation.

Conclusion

Strictly from a power point of view, tracking the sun is clearly worth the effort to gain 28 percent more power from the same cell. The trade-offs include added cost and reliability of adding the stepper motor and the mechanical linkages. Why not just add a second solar panel? The problem is again the limitation on how fast you can charge a given LiPo battery because of how solar LiPo chargers work. The SunAirPlus limits the charge to 1,000mA, for example. If you put more solar panels on the system, you will increase the charge at the lower parts of the curve, but at the peak of the day, it doesn't matter how much power is available to the charger, anything over 1,000mA is wasted, and you will see the solar panel voltage go up as in Figure 2).

This project was not only fun, it was great to see that I could gain pretty close to the theoretical value of power by tracking the sun with some simple hardware and a Raspberry Pi.