In the Spring 2020 issue of the MEPTEC Report, Herb Reiter briefly emerged from his recent retirement to pen an article and interview based on lessons learned from 50 years in the semiconductor industry. Turns out he had a lot to say. So with Ira Feldman and Bette Cooper’s blessing, we’re publishing it as a series. Six lessons each, and a final Q&A. This is the first of three parts is here.
I have been asked to share some of what I learned during my 50 years as a user and supplier of semiconductors and Electronic Design Automation (EDA) tools and methods. Here are brief summaries of my roles and more importantly the major lessons learned in this industry, in chronological order:
After studying electrical engineering for five years at the technical college in Linz, Austria, I graduated as MSEE in the class of 1971. Our school was well respected in the industry, and every graduate had several job opportunities to choose from. I chose the SIEMENS Central Laboratory in Munich as my first employer. My role was to develop the transmit/receive interface card between 120 V cross-country phone lines and an RS232 interface into the T1000, the industry’s first electronic teletype. It received/transmitted “high-speed” ASCII signals (50 bits/sec !?) about 20 miles to a telex exchange and/or other teletypes. YouTube shows how a T1000 looked like and how “fast” it worked in those days:
Lesson 1: It takes many well-coordinated people across a large corporation to develop a complex product, manufacture it cost-effectively in high volumes, and sell it profitably worldwide.
After working with discrete components (transistors, diodes, resistors, inductors, optocouplers…) at SIEMENS for more than two years, I joined a much smaller company in Munich to broaden my scope. I managed manufacturing operations and the development of various data logging systems for telecom, automotive, and marine applications, even for a research project in a nuclear power plant. I liked this versatile role but realized that I lacked knowledge about business topics.
Lesson 2: If you run a small company, you must have development partnerships and a reliable supply chain to succeed.
German law requires that companies share the profits made from inventions with the actual inventors. I registered more than a dozen analog circuit design patents during my work on the T1000 teletype. This product was highly profitable. And in early 1976 SIEMENS sent me a big check for my share of these profits. This large amount of money, tuition-free higher education in Austria, and part-time work allowed me to finance my living expenses during a 4-year MBA program in Linz. Improving my communication, planning, legal, marketing, organizational, and financial skills was a good investment and allowed me to get job interviews at many employers in Germany and Austria.
Lesson 3: Understanding technology is valuable. Complementing engineering knowledge with an understanding of business considerations and human behavior is useful.
In 1980 a young (see photo) and energetic Herb joined National Semiconductor’s European headquarters in Munich as product marketing engineer for a revolutionary speech synthesis chip, branded DigiTalker. Highlights of this job were:
- Business trips across Europe.
- Getting to know Jim Soloman, leader of National’s analog division, and Joe Costello, R&D expert improving speech synthesis and planning speech recognition solutions.
- Spending an entire day with the very intelligent Dagmar Berghoff in a recording studio in Hamburg. She was the main German TV News speaker like Tom Brokaw was in the U.S.
We recorded sample sentences in her voice, with many different paces and intonations, to make them suitable for data compression and storage in a “large” EPROM (16 k bits !? of erasable programmable read-only memory). DigiTalker enabled Ms Berghoff to tell every Audi Quattro luxury car driver when a door was ajar, the gas tank was running low, etc.
How did drivers react to talking chips in 1981? Most of them promised to turn this feature off right away. They reasoned that hearing another voice, in addition to their always nagging spouse, was simply too much. That was the end of the revolutionary DigiTalker and my first experience with marketing ICs.
LESSON 4: Creating a market for a new product is much harder than introducing a good solution that meets current market need.
National assigned me right away to another bleeding-edge technology: Bipolar Programmable Logic Devices (PLDs). I started this role with 20 PLD samples on my desk and no PLD opportunities nor revenues to expand from. It took me, together with in-house as well as external partners, about two years to build an ecosystem for the 14 and 16-pin PLDs in Dual Inline Packages (DIPs). We created data sheets and application notes, qualified several programmer vendors and a design software partner to make PLD customization more user friendly. We trained National’s field application engineers how to use PLDs and they offered training seminars across Europe. All these efforts resulted in PLD revenues of more than $ 1 million per month. However, these happy times did not last. When AMD introduced the faster and lower power CMOS 22V10 with 28 pins and much more logic, the demand for National’s PLDs vanished.
Lesson 5: a) If YOU do not obsolete your legacy products, somebody else will. b) A higher level of integration always wins.
At the end of 1983, National transferred me to its worldwide headquarters in Silicon Valley to market another bleeding-edge technology: Bipolar gate arrays (a.k.a. ASICs). At that time offering faster computers was the primary goal of system companies – regardless of component cost and power dissipation. My ECL gate arrays were a perfect fit. They made money for National, but only for several years. When much lower power CMOS gate arrays became widely available, offered reasonable speed, more pins, and many more gates, I became the official “funeral director for bipolar gate arrays” at National.
Lesson 6: Performance per Watt, not raw speed, is a key success criterion for an IC’s success.
I do not like funerals, so it was time for me to move on. Right after the big earthquake in 1989, I joined the CMOS ASIC technology leader VLSI Technology (a.k.a. VTI) in San Jose as alliance manager for European partners – Philips, ARM, EM, ZMD, and others. My boss, the VP of Strategic Alliances, taught me how complementary and profitable well-managed corporate alliances can be – if they are win-win partnerships.
In part 2 of 12 Lessons Learned from Fifty Years in the Semiconductor Industry – A Memoir, Herb discloses the final six lessons learned. If you can’t wait, you can find the full article in its entirety in the MEPTEC Report.