My family installed our 1,700-watt PV system late in the summer of 2006 (see “Creating a Brighter Future” in HP118). We were immediately impressed with our system’s performance and continued to check actual output versus our projected production each season.
The result of our data-gathering revealed that our annual kWh production was about 10% less than projected. We attributed the energy loss to array shading. In the summer, an overhang partially shades the eastern-most modules until 10 a.m., and a willow tree on the west starts to shade the western-most modules as early as 3 p.m.
In May 2010, in an attempt to reduce the effects of our partial array shading, we installed Tigo Energy Module Maximizers, which have individual module monitoring. And while we are still unsure whether or not the shading problem has been reduced, the value of the module-level monitoring was immediately apparent. Within minutes of adding the units, problems with our Mitsubishi modules No. 2 and No. 4 were identified. Under sunny conditions, the power output from these two was about 22% less than the other modules. Additionally, module No. 8, while not as low as No. 2 and No. 4, seemed to be underperforming as well (about 8% lower). We suspected faulty bypass diodes, but the junction boxes are potted and there was no easy way to test the them.
Because I was having difficulty getting a response from the module manufacturer, I contacted a friend who works in PV manufacturing for advice on troubleshooting the modules. She suggested using an infrared (IR) camera to identify either hot cells or solder bonds or even warm (active) bypass diodes. I located an older, black-and-white IR camera through our local utility. They sent their renewable energy engineer, Jim Heneghan, over one sunny afternoon. Using the camera, he quickly found glowing white spots on all three modules.
In researching common internal module problems and showing the photos to friend and colleague Bill Brooks, I believe that my modules were suffering from poor solder bonds and perhaps shunted cells. I filed the warranty claim, including the IR images and Tigo screen shots as further evidence for the modules’ malfunctions. (Note: In preparation for writing this article, I got color IR shots thanks to a loaner camera from FLIR and HERS rater Gary Handelin.)
The modules have now been replaced under warranty, and are installed in my system. To be safe, I took IR images of the newly installed modules in the array to spot any problems. So far, the replacement modules seem to be working just fine, and no hot spots were detected on the new modules in the array.
It took me awhile to receive a response from the module manufacturer, possibly because, like many manufacturers, they have not had to handle many warranty claims. However, in a few years, the rise of module-level monitoring may change this. While the number of installations with module-level monitoring is rapidly growing, these are primarily with new systems and new modules, and module issues may not arise until after several years of sunlight exposure. At this point, the number of systems with older modules (ours are four years old) and module-level monitoring are very few.
Over the system’s projected 25-year life, the energy loss due to the malfunctioning modules would have equalled about 3,400 kWh—what our entire system typically produces in a year and a half. But I also think about the energy it took for me to chase down the source of the problem, as well as the carbon footprint for returning the old modules and shipping the replacement modules. Considering those factors, I’m not sure if the net energy result is one in our favor or not. But I am pleased with what IR cameras and module-level monitoring have to offer to PV system troubleshooting, and I am much better prepared should we experience future array problems.
Note: In a search for affordable troubleshooting tools, I also used a Black & Decker Heat Detector (about $50) to see if it could identify module “hot spots.” A reference temperature can be set, and if the temperature goes 5.5°C above the reference temperature, the Detector projects a red light. While the Detector does help with troubleshooting, it cannot generate images for use in a warranty claim.