by Ray Zinn
Founder and CEO, MicrelIntuition is perceived to be second sight, clairvoyance, the supposed ability to perceive future or distant events. It is a word tossed about too casually in Silicon Valley, as though we manufacture mediums, not microchips. Entrepreneurial intuition does exist, but it isn't magical. It is driven by the relentless observation of everything. It is how I came to conceive of a product that is now a standard part of the silicon chip-making industry and sell it to Texas Instruments before I ever wrote an engineering specification for it. It also led me to start Micrel because the entrepreneurial drive made me an ill-fitting cog in the machinery of slower-thinking companies.
I conceptualized a machine called the wafer stepper. Without going into the technical details, it was an intuitive evolution in the way the microchip industry uses light to create the semiconductors in everything from a refrigerator to an iPhone. Without the wafer stepper, shrinking transistors to their current size might not have been possible or practical. At the time, silicon etching technology allowed for accuracies of around 1 micron, or one-millionth of a meter, wide. Today the state of the art in mass-manufactured chips is 14 nanometers, almost 0.00000001 of a meter, or 100,000,000 times smaller than the standard of the time when I conceptualized the technology that has brought us to today's ever-shrinking silicon footprint.
The old process was based on using light to create photomasks: templates that filtered light during the etching of silicon wafers to create millions of transistors on a small piece of silicon. I got the idea not to use photomask-making equipment to create these masks but to image the transistors directly onto the wafer and eliminate mask making altogether. The benefits were enormous. The density of transistors could be radically increased, and the technique could easily be adapted to different wafer sizes. Accuracy and chip quality would also improve.
The problem was that it didn't exist outside my mind, and in that era nobody would want it because there was room to shrink chips by using the prevailing technology. But years later, when Apple put a computer many times more powerful than the Space Shuttle's onboard flight computers into your pocket, wafer steppers would be essential.
When I dreamed up the wafer stepper, the company I worked for was in bed with IBM, pushing a chip-making technology called e-beam, short for electron beam. Ours was a small company and had to ride this big horse while it could; there were no spare resources for new concepts developed by a salesman, which I was at that time. So the concept stayed in my head as I was sent to make Texas Instruments (TI), one of the enduring global chip leaders, a customer.
I was not immediately successful. I was turned away once. On a second try at introducing myself to a key TI contact, I was physically ejected from the building -- tossed to the curb by two strong fellows, ripping my pants upon landing. My boss was encouraging in the old Theory X style of management, letting me know that failure to get the TI account was an employment-limiting prospect. "Any good sales guy will find a way to get in" was his lightly veiled ultimatum.
But TI was not interested in our regular photomask equipment, instead buying its gear from other, larger vendors. After my second attempt, my boss said, "You got to go back. It's important we get in there. You got to make some inroads." I couldn't think of any way to do that, but I had been thinking a lot about my wafer stepper idea and how it could in theory improve chip geometries. In those seemingly ancient days, the wafer stepper idea could shrink geometries down to 435 nanometers (0.435 micron), or about a third of what traditional photomasks could manage. Most people, though, would have asked why anyone would create a circuit for 435 nanometers if you needed it only for 1.5 microns. It was a technology to create devices nobody yet needed.
But I had a hunch that TI was already thinking beyond the limitations of current state-of-the-art wafer etching equipment, and so I devised a scheme to at least get a fair hearing for our traditional gear and then insert the wafer stepper concept into TI's collective engineering mind.
The third time I went to Texas, the receptionist asked me who I wanted to see and who I was. "Tell him his brother is out here," I told the receptionist, and then sat on the waiting room sofa. After a brief moment, I saw my soon-to-be customer coming down the stairs. He looked around for his brother and, not seeing him anywhere, asked the receptionist. She pointed at me, and when he glanced in my direction, I put a Cheshire grin on my face. He stormed over, obviously madder than it is safe for any Texan to be. "How dare you tell the receptionist that my brother is here to see me?"
"Well, aren’t we all brothers and sisters in some way?" I said in the friendliest tone I could manage. He broke down and started laughing loudly. He picked me up by the arm and said, "Okay. Come on."
Once we were in his office, I started talking about the stepper. Keep in mind that at this stage of the game, the stepper existed only in my head. My company didn't know about this, and there were no specifications, no engineering plans, no cost estimates... nothing. I wasn't even expecting this engineering manager at TI to want the stepper. I just wanted him to believe that we were thinking long-term and thus that buying our existing equipment and building a relationship with us was a good idea.
But he was interested in the wafer stepper and called a couple of other people in to hear my story. I was very convincing about the nonexistent device because I thought it was a great idea but mainly because I needed TI as a customer. I did a little dog and pony show, making diagrams on the fly while telling them that the vaporous equipment was top-secret and that Texas Instruments was very lucky to be hearing about it so early. I certainly did not want them to
go back to my company and tell them anything about wafer stepper equipment because it didn't exist and it was contrary to our being embedded with IBM on the e-beam technology.
The problem was that Texas Instruments was interested. They understood the benefits of the wafer stepper concept, saw value in it, and thus didn't want to discuss our actual, existing, purchasable equipment. They simply would not stop pushing me for details about wafer steppers. They disclosed that they had plans to shoot for half-micron chips because they were actively manufacturing microwave transistors and needed wafer stepper–level optical capabilities. The more I tried to discuss equipment I could actually sell to them, the more they wanted to discuss equipment that existed nowhere.
I went back home and never uttered a word about the wafer stepper except to my boss, who would want details about my foray into TI. Indeed, he specifically wanted to know if Texas Instruments was going to order any equipment, and so I said without a lot of confidence, "They’re thinking about it. I gave them some literature, and they're going to get back to me."
Meanwhile, word of my wafer stepper idea circulated beyond my manager's office. Of course nobody else in the company had any idea what a wafer stepper was, so they called me. "Oh, it's an idea I was just telling TI about," I told them, and so my management did not take any action on corresponding with TI since they were focused on our IBM relationship and their e-beam technologies. Back office communications quickly died down, as I had hoped, and I didn't think more about it.
But TI did get back to me, wanting not to place an order for equipment we made but to ask when they could expect a quote for wafer steppers. I asked, "Well, how about my other equipment? Is there any way we can talk about that?" They were marginally willing to discuss our existing gear but again pushed for a quote on the wafer stepper.
This put me in a bit of a bind. How does one create a quote for something that their own engineering department had never spent a moment pondering, much less designing and building? I couldn't let TI go. Getting them a quote for wafer steppers was key to getting them to buy our existing products. Therefore, I decided to highball Texas Instruments. I fetched some quotation paper and typed up a quote on my own. Given the radically new capabilities of the equipment we had yet to invent, I decided that a price 10 times the going rate would be high enough to make TI blanch. At $800,000 a unit, even TI would not put such equipment into production, for it would cripple their profitability.
Never underestimate Texans or the power of market momentum.
They didn’t blink. Not only was the $800,000 price not too high for them, they wanted four machines at that price, for a purchase of more than 3 million bucks, which was huge in the era of 1-micron semiconductors. The purchase order arrived with an urgent request for a delivery date for equipment my company didn't even know about.
My big mistake -- though in the long run it wasn't a mistake at all -- was to actually create a quote. Once issued, the purchase order bordered on an agreement. This put my and my company's reputation in jeopardy. Soon Texas Instruments was calling my company's headquarters, wanting to know about the status of its order. Soon my phone was ringing: "These Texans are asking about something we have never heard about." I rattled off some specifications over the phone. The somewhat bewildered voice from HQ said, "This hasn’t been seen by engineering!"
Creating specifications on the fly didn't work, and soon enough things got so out of hand that my boss finally called me down to the Woodland Hills home office. "What is this wafer stepper issue that TI keeps calling us about?" I gave him the background on the idea. "Well, that is not a trivial thing," he complained. I told him that TI would pay $800,000 for such a machine, and that got my boss's attention. "Oh! That’s not bad. But given the up-front engineering work, I wouldn't even consider doing this unless we got an order for five."
I reached into my coat pocket and pulled out the Texas Instruments purchase order. "Would you accept four?" I think he quit breathing for a minute or two.
We accepted the order and started designing the machine. This required diverting resources from our e-beam project with IBM. But if TI was eager to buy four borderline fictitious wafer steppers at $800,000 a pop, maybe IBM would be interested in the technology as well. We sent the TI quotation and specs to IBM.
IBM's vice president of research wrote us back a stinging letter saying that the wafer stepper was the dumbest idea ever voiced in the entire chip-making industry (keep in mind that it is now the de facto method for etching chips). This was a huge problem since IBM was our biggest and most important customer. My boss tossed IBM's letter across the desk to me, and after reading it, I turned pale white. IBM's head of R&D was the most important industry scientist at the time, our company was insignificant, and the wafer stepper was little more than notepad scribbles.
"You really shouldn't work for anybody else," was how my boss summarized the fiasco/opportunity my wafer stepper idea created. "You just don't fit in. You're not the kind of personality that belongs in an organization. You really have to work for yourself." It was an invitation to "pursue other endeavors." It would be six years later that Tom Peters would write in his book In Search of Excellence, "If you’re not getting fired, you're just not trying hard enough." I certainly had tried hard enough, and I did get fired so to speak. We both agreed it was time for me to leave.
I was walking up the stairs from the garage to my house. My wife met me at the stairs, as she normally does. "I've quit my job," I announced. "I'm no longer going to work for anybody. I'm just going to work for myself." It was 1976, a year when the nation was experiencing stagflation and economic uncertainty abounded. "Oh, okay," was her reassuring reply. "What are you going to do?"
"I don’t know. I'll figure it out." And with that bit of confidence, I collected my last paycheck. I brought Micrel Corporation to life two and a half years later, in November 1978. It took a few steps in between because I didn't have any idea what I was going to do other than that I knew I was not working for anybody else.
Funny thing, though. E-beam etching technology basically doesn't exist today, and the wafer stepper is one of the most important pieces of equipment in our industry. The brilliant IBM vice president of R&D was simply wrong, being so thoroughly wed to e-beam that he could not accept a different approach. Like most innovations, the wafer stepper started with a bit of informed convergence, which is the essence of entrepreneurial intuition: seeing enough of the pieces of a puzzle that you suddenly see how the puzzle can be pieced together. Steven Jobs understood calligraphy and computers and saw ways to make one do the other. I saw how existing photomasks were limiting transistor shrinkage but how the basic photographic setup could be simplified to increase chip density.
Intuition is related to wisdom. Wisdom is the proper application of knowledge. We all gain knowledge, but if you lack wisdom, it means you will fail to see how to apply that knowledge for the improvement of anything, whether chip design or humankind.
*********************************************
This article is an excerpt from the book Tough Things First (2015) by Ray Zinn and is published with permission by McGraw-Hill Education.
Copyright 2015 MEMS Journal, Inc.