Start • System Overview • Applications • Hardware • Power Supply • Photosensor • Program • Improvements • Sources and PDF

On a seemingly rare sunny day in Ithaca, New York, the sun delivers about 1,000 W of power per square meter and just begs to be put to some purpose besides browning the backs of the students lying out in the gorges. At up to 15% conversion efficiency, COTS photovoltaics (PVs) can turn that light into electricity. A growing number of Ithacans have heeded the sun’s call and installed solar electric PV systems to power their homes and businesses. Ithaca may get 40% less solar insolation than San Diego, but it gets 25% more than Germany, the world leader in installed PV capacity. Tompkins County, where Ithaca is located, has roughly 2.9 W of installed solar capacity per person, which makes it second in the U.S. to only Palo Alto, California.

With the Finger Lakes, ridgelines, and valleys cutting through the region, clouds and fog levels vary significantly from one part of town to another. But weather data is only available from a few select locations. How is a potential PV buyer to know how much power her system will produce unless they can measure the incoming light? And how can a proud PV owner know how her system is performing or detect faults unless she can confirm the conversion efficiency? Expensive commercial data logging systems that cost thousands of dollars do this well, but they are entirely unreasonable for the small system owner.

For Bruce Land’s ECE476 Microcontroller Design course at Cornell University (http://instruct1.cit.cornell.edu/courses/ee476/), I designed an inexpensive self-powered solar data logger to meet this need (see Photo 1). I built the system around an Atmel AVR STK500 development board that featured an ATmega32 microcontroller. You can leave the logger (untouched and isolated) in the field to collect data for months or years.