Cree in the New Forbes
Imshaken, Become a Complete Fool
I know that the quote I highlighted is a paraphrase of Swoboda by the Forbes magazine writer. But then the writer quotes Swoboda:
"There's a little bit of art to this science. Nobody really knows why things work," says Charles Swoboda, 35, Cree's chief executive. "There are theories, that's all. We know what works, but not how it works."
The semiconductor industry, in my opinion, has progressed at the rate of Moore's law not out of a lack of understanding of what they were doing but rather because of a necessity to wait for the necessary infrastructure to appear: the reactors, the lithography, the cleanliness etc. and whatever else is needed to go to finer and finer line widths. It is not slow because of a lack of understanding of the basic processes involved.
All along the way they understood where most of their hurdles would be even ten years down the road. They understand this so well that they keep a running road map of their proposed technology goals for each of the next three to five process cycles. All along the way they keep track of missing technology that will be necessary four to eight years down the road. The purpose of this is so that they can plan for filling in those gaps ahead of time through their research programs. And if some of that research is not bearing fruit as the drop dead date for it nears, there are well defined alternative, though less desirable, paths for the industry to take and still get to the goals for that cycle.
If, for replacing incandescent lighting, Cree really needs to get to something beyond 50 mW per LED (with a theoretical max of 70mW representing 100% efficiency) and they do not have concrete ideas of the steps they need to take, the problems they have to solve in order to get there, then it is no longer an engineering job but artistry that they are depending upon to get there. Artistry does not come with the kinds of guarantees that engineering does. There may be unknown scientific barriers that will forever prevent this.
For example consider this: the present 15 mW XBrights are about 21% efficient given that the theoretical limit of 70 mW, (3.5 volts times 20 milliamps), represents 100% efficiency. That means Xbrights somehow waste 79% of the possible photons that could possibly be produced by the applied power. Do we understand if all the input power generates primary photons but 79% of those photons never get out of the crystal? -- because the primary photons get reabsorbed and thus get converted to secondary heat? Or does some of the electrical energy merely create primary heat along with the primary photons? -- heat will never be reclaimed as light.
I don't know what this ratio is, (primary photons to primary heat plus primary photons), but I sure would hope they do know this in the labs. And I hope it amounts to at least 50 mW of primary photons, for 70 mW applied electrical power. If not the desired efficiency will be impossible even if all the primary photons are somehow coaxed out of the crystal.
Ever since Edison, we have played with the incandescent bulb. It will never become more efficient than it is now, about 10%, because the very process by which it converts electricity to light creates nine times as much heat energy (in the form of infrared photons plus resistive losses conducted through the hot glass and metal base) as it does visible light photon energy. Engineers have worked on this for ~120 years now. There are no more engineering improvements that can reasonably be made. The artistry needed to go to further efficiency leads one to fluorescent bulbs instead.
Not trying to scare everyone. A company can thrive without replacing all the incandescent bulbs in the world. And it is still possible to replace all the incandescents even with the current XBright efficiency (21% for XBrights vs about 10% for incandescents), provided the price of the XBrights can continue to drop. That price drop may be possible with engineering alone rather than the necessity to count upon artistry.
It's just that the lack of a true long range engineering roadmap, which is what Swoboda's statement really implies to me, means that he is speaking out of faith, not knowledge, when he predicts a ten year time frame for competing with incandescent lighting. That is what scares me. We saw what Cree's revenue estimates last year, based upon faith, did for the price of the stock.
Join the best community on the web! Becoming a full member of the Fool Community is easy, takes just a minute, and is very inexpensive.
Become a Complete Fool