The warranty chain trap

The installation company, whose director responded to our call for "black sheep," has installed around 2,500 photovoltaic systems, according to information provided by the company. But this case could have ended the company’s success story; it was an uncomfortably close call. The installer, who is from North Rhine-Westphalia, reported that his company installed a PV plant with a rated output of 175 kilowatts and a total net value of around EUR 700,000 euros, for an end customer in late 2008. A wholesaler the company had been working with for many years delivered the entire system, including the modules, racks and inverters. The installation went smoothly. Twelve months later, however, the customer complained of significantly lower performance than in other systems he operated. The negative deviation amounted to roughly 20 percent. "At first, we looked to see if we had made a mistake," said the installer. "But after I had gone out five or six times and checked everything thoroughly, it became clear that it wasn’t an installation problem, but that the modules were to blame for the poor performance."

According to the manufacturer’s flasher list, the average module output should have been approximately four watts above the listed rated output of 150 watts. An expert measured the performance curves, according to the installer, and found reduced performance of up to 25 percent. A test of two modules at Fraunhofer ISE revealed that both modules performed at a level 20 percent below the specified capacity in a flash test. Taking into account the output tolerance of plus/minus five percent and measurement inaccuracy of plus/minus three percent, the modules still demonstrated clearly reduced output of at least 14 or 15 percent.

Goodwill was not enough

Based on these findings, the installer hoped at first that he could fall back on the dealer’s guarantee or the performance warranty of the manufacturer. The performance did deviate substantially from the assurances made on the data sheet. But nothing came of it. The installer’s warranty to the end customer was two years, but the dealer’s warranty to the installer was only one year, which had just expired. The installer was stuck.

At first the dealer offered a goodwill compensation of nearly 70,000 euros. However, the end customer rejected this, saying it was too low. The installer wanted to try negotiating with the dealer for a higher amount that reflected the average reduced performance of the modules. But once the dealer was confident that he could not be legally held liable in the case, he revoked the original offer and refused to pay any compensation, said the installer. "At the same time, the end customer’s lawyer began to put on pressure by setting a deadline for the compensation and threatening a lawsuit if the compensation wasn’t forthcoming."

Trial without support

The trial against the installer began. "I was getting a letter every other week from the end customer’s lawyer or the expert evaluator or the court. They put me under constant pressure. Shortly before the verdict was announced, both the dealer and the importer told me that it was my problem and that I was on my own. I couldn’t expect any more support from them. Ultimately, that’s what happened. I didn’t have a chance."

A series of unfortunate circumstances came together in the installer’s case. For one, the reduced output of the system was only discovered after the dealer’s warranty had expired. This was partially due to the fact that the installation of the monitoring system was delayed for nine months because the chicken coop on which the system was located had no internet connection. There was no management contract between the installer and operator.

When the monitoring finally worked, it was already October. However, fall precipitation and insolation conditions made it difficult to draw conclusions about the performance of the system based on its yield data. This can only be done reliably if the monitoring takes into account insolation values, i.e. data from an independent insolation sensor or external weather data. This was not possible with the simple monitoring system installed in the plant, which meant that no conclusions about the performance ratio of the system could be made. Even an undesirable drop in performance was not evident from the monitoring, because the modules’ capacity was lower from the start. The operator only noticed that something was wrong with the system because he compared the output with that of other systems nearby.

Poor warranty conditions

The second problem arose from the warranty conditions of the module manufacturer. Only 80 percent performance was guaranteed after 25 years. The manufacturer’s warranty at that time did not guarantee, for example, 90 percent performance after ten years of operation, which is now common. That means that even if output is 20 percent lower from day one, it is still acceptable under these warranty conditions. Additionally, the modules were stated to have an output tolerance of plus/minus five percent. "If you assume a certain level of measurement uncertainty, the manufacturer or importer could argue that reduced output of more than 25 percent was normal," said the installer.

Ultimately the installer paid the end customer not only damages, but also for many court and attorney fees, adding up to approximately EUR 175,000 euros. "We paid the last installment of the damages to our end customer in March of this year," said the installer. "The worst part of it was that our nerves were totally shot." It was scarcely possible to manage the company properly throughout the seven years. Only his responsibility towards his employees motivated him to keep the company running.

Trust versus control

The installer felt left in the lurch by the law in this case. He claims that even the judges shook their heads at the end of the trial and indicated that they found the solution unfortunate. Ultimately it was the gap in the warranty chain that condemned the installer to his fate. The installer believes that, in order to prevent such cases from arising, legislation is needed to ensure that the same warranty period that applies to end customers also applies to dealers.

In his opinion, this warranty gap is still common for most installers. He knows of multiple cases of other installers who have experienced similar situations. "If the manufacturer only offers the dealer a warranty period of one year, then I can hardly convince the dealer to give me two years’ warranty," he said. "Otherwise the dealer would be stuck sitting on the warranty for a year." A dealer did once offer to extend the warranty – for a significant price. But it was much too expensive.

In retrospect, the installer admits one mistake: "I should have checked the products right after I received them." Instead, he simply trusted his dealer, with whom he had worked for many years. He will not make a mistake like that again. "I still work in the solar business, but I’ve become much more cautious." And if something similar should somehow happen to him again, he would go straight to bankruptcy court and give up everything, he said. "I never want to go through a circus like that again."

The dealer involved did not comment on this case by the time this issue went to press.

Questions for discussion at our round table at Intersolar:

· How can an installation company check the output of delivered modules at an acceptable cost?

· Is the performance warranty often misunderstood and does it give a false sense of security?

· How should contracts be evaluated in which the warranty period legally applicable to the end customer is shorter between the dealer and installer?

· Can deficient modules with reduced performance be compensated for by adding modules, i.e. by expanding the system?