Your daily dose of Photovoltaic Technology Developments and Solar News

News Feeds | Newsletter | Advertise | About Us | Subscriptions

Order your copy of Photovoltaics International Edition 1#

ASU’s Photovoltaic Testing Lab shakes, bakes, zaps modules in the cause of safety and reliability

14 May 2008 | By Tom Cheyney | Chip Shots

As we walked through the dusty, weedy area of previously tested solar modules known as the "boneyard" at the Photovoltaic Testing Laboratory (PTL), my tour guide and lab marketing manager Paul Symanski warned me that "there might be snakes" among the piles of Solar Power Corp., Astropower, and other museum-worthy units. He offered this comment nonchalantly, as if he were asking me if I took cream and sugar in my coffee.

The PTL occupies about 2.5 acres of Arizona State University's East or Polytechnic Campus, in a part of Mesa outside of Phoenix that blends living desert and increasing suburban sprawl. Snakes, venomous and non, come with the territory, especially when temperatures start to warm up and the creepy crawlies come out of hibernation in the spring. And it can get pretty toasty out there in a lot jammed full of PV panels.

PTL's mission: To test and to train.
(Photos courtesy: PTL)

No rattlesnakes slithered out of the stacks of aging solar gear or anywhere else in PTL's yard during my visit, but I did see hundreds of south-facing silicon- and thin film-based modules and concentrator PV arrays undergoing safety, reliability, and other performance evaluation/engineering tests. As the only independent ISO- and IEC-accredited US lab for design qualification and type approval testing as well as a training ground for solar industry professionals (half the workers are students), PTL has been very busy of late. Symanski said that business has quadrupled over the past two years, with the majority of that growth in the last year or so--and expansion of the facilities has been proceeding at a furious pace.

When I toured the site in late April, Symanski and William Shisler, PTL's quality manager, told me that there were 87 concurrent projects running, with about eight modules per project, averaging four to five months of testing per project. Since its inception in 1992 (it became accredited in 1997), the lab has tested more than 3000 different types of modules. Some 90% of the modules tested are of the mainstream crystalline-silicon variety, with the remaining 10% comprised of various thin-film types, although the lab expects a higher percentage of TFPV moving forward. Rigid and flexible, flat-roof and building integrated, stand-alone and grid-tied--PTL's module farm has just about any type, shape, and size you can think of.

Nothing like desert sun to test a solar module.

The gear is subjected to a wide and challenging range of mechanical, electrical, and environmental simulation testing. The modules are heated up, cooled down, soaked, pounded, shaken, beamed, zapped, and subjected to accelerated aging. Of course, as Shisler pointed out, the most impressive testing tool is that ubiquitous Arizona desert sunshine, the same single point source beaming at the same level, providing what he calls "the biggest solar simulator in the world." But the large thermal cycling and custom-built UV chambers, laser-guided catapult ice cannon, impact test bed (with its hundred-pound bag of lead shot), and new high-resolution CPV trackers are pretty impressive too.

PTL shoots these iceballs at modules
to see if they can survive a hailstorm.

As long as the modules' frames don't collapse or glass fragments don't damage the cells, the they can get pretty beat up--with dents and cracks and such--and still pass the rigorous testing. Although many modules in the yard have circles that have been drawn around dings and scratches during visual inspections, those rarely translate into system killers. Moisture remains the biggest failure mechanism, as Symanski reminded me, so sunny places with high humidity or proximity to bodies of water (like Florida) can be especially challenging environments for modules to operate reliably over a long period of time.

Shisler explained that "you can't bake something straight out of the box and think that's how it will operate." As examples, he noted that a single-junction amorphous-silicon thin-film module degrades signficantly with first exposure to the sun before it stabilizes and can then be measured for its true conversion efficiency, for example, while a copper-indium-gallium (di) selenide (CIGS) device performs better at first exposure, then stablizes to a more accurate efficiency level. Some manufacturers' modules with identical cells and nameplates may not have been fabricated in the same way, resulting in differing levels of conversion efficiency and durability, something PTL is adept at measuring.

In a paper to be presented later this week at the IEEE Photovoltaic Specialists Conference in San Diego, Shisler and Symanski are among the coauthors of a study led by PTL director Govindasamy Tamizh-Mani comparing failure analysis rates of design qualification tests. The report reveals a surprising failure rate trend: a large increase in fails was seen in the 2005-2007 period compared to 1997-2005 time-frame, for both c-Si and thin-film modules.

Across the board, the more recently conducted accelerated tests like damp heat, thermal cycling, and humidity freeze showed a higher failure rate than the earlier tests. For c-Si modules, the biggest change was found in the diode thermal tests (29% fails vs. 8% before), while for the thin-film panels tested, the damp heat evaluation had the most pronounced change (70% vs. 28%). Of the 31 different manufacturers' modules that underwent bypass diode testing, 19 consistently passed, 2 consistently failed, and 10 randomly passed the tests.

Since customer confidentiality is one of PTL's cornerstone policies, don't expect to find a consumer reports-type listing of the companies whose modules--good, bad, or inconsistent--were tested during the study. But do expect PTL to keep up its relentless barrage of testing, to make sure the modules are safe and perform as the manufacturer intended--and the consumer demands.