Solar Cells Info

by Tom Cheyney, PVTech, 25 May 2009

http://www.pv-tech.org/chip_shots/_a/an_interview_with_first_solars_bruce_sohn_part_I/

At First Solar’s corporate headquarters in Tempe, Arizona, a morale-boosting slogan adorns posters stuck to the outside of cubicle partitions: “MILESTONE MADE! TEN ONE ONE.” That’s “Ten,” for 10 years in business–at least in the company’s First Solar incarnation. The original firm Glasstech Solar, led by visionary Harold McMaster, actually set up shop in 1984, then became Solar Cells, Inc. in 1992, which begat the present entity in 1999. The middle “One” stands for the gigawatt’s worth of panels produced in the solar module factories in Ohio, Germany, and Malaysia – as well as the annual production capacity that will be ramped by the end of 2009. The final “One” stands for perhaps the biggest accomplishment of all–the dollar-per-manufactured-watt standard beaten by two cents by First Solar in the final quarter of 2008, a cost that has since shrunk to 93 cents per watt in the first quarter of 2009. But then, “Ten/One/0.93” doesn’t quite have the same ring.

When Bruce Sohn joined First Solar as a member of its board in 2003 and then full-time as its President in March 2007, that Ten/One/One level of achievement may have seemed a distant goal as endowed as much with hope (and maybe a few prayers) as realistic expectation. A comparison of the company’s financial and manufacturing performance at his hire date and in the most recent quarter shows a staggering growth rate–in dollars, watts, and production efficiencies.

Revenues for the first quarter of 2007 were US$66.9 million, with net income a promising US$5 million. For the first quarter of 2009, those same amounts were US$418.2 million and US$164.6 million; in other words, the income was more than double the sales from just two years before. During the same period, gross margins grew from the mid-40s to the mid-50s.

As for solar power products coming off the manufacturing line, tens of thousands of cadmium telluride (CdTe) thin-film PV panels have become millions of modules per quarter, with megawatt production levels jumping from the twenties to the two hundreds. During that time, the annualized capacity per line (the proverbial run rate) jumped from the mid-30s of megawatts to 49.4MW, and conversion efficiencies have popped a full percentage point, from 9.9% to 10.9%, with a roadmap for several more digits of improvement.

Sohn, a 24-year veteran of Intel and its ultrafocused manufacturing strategies, has played a crucial role during First Solar’s prodigious yet well-managed, tightly controlled growth spurt. Deftly handling a laundry list of responsibilities, the Massachusetts Institute of Technology engineering graduate is charged with overseeing the thin-film PV juggernaut’s operations for its module business, including technology development, manufacturing  and expansion; its systems engineering, procurement and construction (EPC) group; and monitoring and maintenance activities.

In an exclusive interview, Sohn talks about the challenges of his job and how it has changed over his two-plus years in the executive cubicle (no suites or corner offices at First Solar). He discusses how his experiences at Intel, where he helped hone the chipmaker’s “copy exactly” factory replication strategy, have been “mapped” to the PV company, including its own “copy smart” approach to cloning manufacturing facilities–and even its engineering, procurement and construction arm.

Other topics include more details on the company’s manufacturing practices, the need (or lack thereof) for standards, the growing role of the EPC group in the U.S. market, the recent acquisition of OptiSolar’s project development pipeline, and the nature of the competitive landscape.

No such thing as a typical day
Sohn laughs when asked what his “typical day” is like, as there is no such thing. He spends about half his time on the road, visiting the company’s factories and development team offices, as well as customers, who are mostly in Europe. When he’s back at the company’s sixth-floor headquarters, he keeps tabs on his “direct reports” and operations via a combination of web-based tools and phone calls, while also closely collaborating with fellow First execs such as marketing/biz-dev maven John Carrington, money-man Jens Meyerhoff, and of course, the boss, CEO/chairman–and soon to be ex-CEO and executive chairman–Mike Ahearn. Most of the time, the big strategic picture weighs heaviest on Sohn’s mind and task plate.

“In any particular day, I’ll probably spend a fairly small amount on actually doing operational things,” he says. “Whether it’s talking about a customer development or working with supply chain management, for example, or future technology work and how that’s going to integrate into the business, I would say most of what I work on is out farther in the future.”

The former Intel fab manager still has a fondness for the operational side. “As an old manufacturing guy, I find interest in all parts of the business but the reality is that I leave the day-to-day to the people who are more in tune and expert with the day-to-day, but I still find a fascination with it.”

Scaling production, developing ‘copy smart’
When he first joined the company in early 2007, his focus was squarely on how to scale the company’s manufacturing and set up the internal business systems and processes to make that happen successfully.  “When I came in, the focus was on starting the replication process,” Sohn recalls. “We had not yet started up the factory in Germany. It was in the latter stages of construction and moving into the early start-up phases. There was a lot of focus on bringing up that first factory, and we spent a lot of time talking about what copy smart was for us, and how we apply that.”

As he admits that he still has to catch himself from saying “copy exactly,” Sohn explains the close relationship between the two factory-replication schemes. “Copy exactly and mapping that over into First Solar and using what we call copy smart, there’s really not a lot of difference between the two other than the name,” he notes. “There are some things that Intel can afford to do that we don’t necessarily endeavour to do because Intel turns over its technology every two years, and we don’t necessarily, since it’s more of a continuous process and we approach it somewhat differently. Nobody should walk away thinking that copy smart is any better than copy exactly. It’s not. It’s just the application of the same basic approaches to a different industry.”

Sohn continues his recollections of his first year on the job. “Back then the company was fairly small: we only had the three lines going, 60MW shipped in 2006, so we had to grow dramatically. That changed our scaling focus and then we’ve also expanded in the way we approach our technology and putting in roadmaps and things along those lines.  We created more detailed roadmaps not only for the technology and efficiency but also for the cost reduction, which is core to our mission.

“There’s been a lot of development of the business, putting in processes, putting in systems, making sure that we could scale,” he continues. “The last thing I wanted to see was that we come in and we start to grow the business and we collapse because we don’t have the processes in place, we don’t have the systems in place, we don’t have the culture in place, or we haven’t enough people or the right kind of people that know how to manage and run a larger business.”

As Sohn recalls, the company’s approach to factory and production-line sizing and site location has evolved significantly. “At that time, we thought we would replicate much smaller 25MW facilities in locations all over the world. Then as we got into the cost and understood how we needed to scale the business and scale the costs over a larger volume, it became clear that we needed larger factories.

“In each of those factories at that time, a line was designed for about 25MW nominally and today those same lines are pumping out more than 49MW, so we’ve been able to almost double the output per line over the last two years as well, which is very important because in essence that saved the need to build some 20 lines or so.”

“If we actually had to build those 20 lines, that’s five factories–I’m not sure that we could have done it in these two years, even if we had wanted to,” Sohn admits. “On the other hand, if we improve the efficiency, or we improve the run rate, or we improve the yield inside the factory, and then we propagate it using the copy smart philosophy, then we see it propagate at 20x the rate, or 23x the rate, because we have 23 lines running.”